"So, hello and welcome to Sartoria's webinar titled: 'SartoBind Break Free from the Protein A Low Productivity Paradigm' presented by Ricardo Bousser and Geoffrey Papersak. My name is David Johnson and I'll be your moderator today. Let me now introduce the speakers that are going to be giving you the webinar today. Okay, so Geoffrey on the right there Geoffrey Pressac is Manager of South Europe for the chromatography field application specialists at Sartorius, supporting customers on a wide range of topics including process development, design of experiments and scale up, plus also training and troubleshooting on our systems.

Geoffrey has a wide experience in purifying both proteins and large molecules like exosomes, plasmid DNA and a wide range of viruses.

Doctor. Ricardo Boussa joined Sartorius back in February twenty eighteen as a product manager for membrane chromatography. Prior to joining Sartorius, she worked as product and marketing manager for affinity chromatography solutions used for recombinant proteins. She has a PhD in biology and biochemistry from the George August University of Guttingen.

During her time as a doctoral candidate at the Max Planck Institute of Biophysical Chemistry Gutteringen, she worked on upstream and downstream process optimization of recombinant proteins from bacterial, mammalian and insect cell cultures. Ricardo also holds an MBA from the European Fernhorschule Hamburg in general management, where she specialized in digital and international marketing. So, with the introductions over, we've got a great webinar for you today. Let's get started and we'll be back at the end to answer your questions.

So, Jeffrey and Ricardo, over to you.

Thank you very much, David, for the kind introduction. Also, welcome from my side to this webinar.

And at the very beginning of this webinar, I would like to give you a general introduction to membrane chromatography before we jump into the other agenda points.

You probably know that there are different formats of chromatography media out there in the market. We have, for example, resins, which have an average pore size of fifteen to two hundred nanometer, And they are mainly driven or their mass transport is mainly driven by diffuser flow.

And resins are typically packed into columns, so there is packing required.

And to make them economically viable, you will or in the industry, it's very often used over multiple batches. So they are reused over those batches.

And to reach a certain dynamic binding capacity, the average residence time for most of the resins out there in the market ranges between two to six minutes.

So let's say, the average residence time is four minutes for a typical resin. On the other side, we have membranes. Membranes have much larger pore sizes, typically between four to six micrometer when we talk about the zaatubius membranes. The mass transport is mainly driven by convection.

They are provided ready to use in our typical capsules or cassette formats, so there is no packing required. And mostly they are handed as single use or single batch use consumables, so they are discarded after using them over one or after one batch.

The typical residence time for bind and elute steps of those membranes ranges between six to twelve seconds.

The major benefit of membrane chromatography is that over a wide range different flow rates, they keep binding capacity constant. So they have a constant binding capacity over a wide range of flow rates. On the other side, when we talk about resins, they have a very high binding capacity at short or at low flow rates. But when you increase the flow rate, the binding capacity very quickly drops. And that makes membrane chromatography very interesting to look into those when we want to read higher flow rates.

So again, typical resins, where the market is used to it, they are typically provided or packed into traditional glass columns. So there is a lot of hands on time associated to packing those columns. You will have to reuse them over multiple batches to make them economically viable because they suffer from low productivity typically.

Or they provide a low productivity only. To get around the high hands on time, there are prepack columns also out there in the market. But even though those have to be reused over multiple batches to make them economically viable, and they still have or can provide only low productivity.

On the other side, have here the membrane absorbers. As mentioned earlier, they are provided ready to use in certain devices like cup pools or cassettes.

And due to their high productivity, they can be fully utilized within one batch. So you can use them, to their full lifetime extension within one batch and then simply discard them so that you get a round of extensive cleaning, validation, storage and so on.

Membrane absorbers are already well established in the market, especially for flow through polishing, of monoclonal antibodies, but also for capture of large entities like viruses.

And when they have been used for flow through polishing, for example, we have seen in the past that a five liter jumbo capsule, which you see here on the right side, can replace a two hundred liter chromatography column, if you want to reach the same volumetric flow rate with the column as you can do with a membrane absorber. And this is because the flow rates for membrane absorbers are typically ten to twenty times higher compared to a standard resin. And this will allow you to decrease process time of up to seventy five percent in flow through polishing, use up to seventy five percent less buffer. There is no cleaning validation required because you will discard them after using them within one bed. And also the membranes have been shown that they can provide comparable virus clearance, which is comparable to dose of resins.

And here I would like to hand over to Geoffrey.

Thanks a lot, Ricardo. Thanks also, David, for the kind introduction.

So now, after a quick introduction on the technology itself, I'd like to highlight how membranes can actually help to intensify the process.

If it works.

Okay.

So first thing I'd like to do is to explain the structure of our membrane. So on the left, you can see traditional chromatography using resins. You can see bleeds that are packed together.

These are usually in the range of fifty to one hundred microns. You can see the target molecule in green here that wants to access ligands in black. And for that, they often have to diffuse for a while. Let's say in the range of ten micron depth usually it can be more can be less.

Because of that, you need what we call diffusive flow. So that means basically a high residence time and a gentle flow. If you go to a convective structure, which is like the traditional membrane chromatography on the market. So you have all sizes which are quite big.

So as Ricardo said, it's often more than three microns. It depends on the membranes. The basic principle is always the same. You have a solid structure in gray with ligands that are grafted on the surface.

Sometimes there's also a hydrogel on the top.

And you have your binding molecule that access for ligands easily because they are all directly accessible. There's no diffusion depth.

Of course, you have a much lower capacity than traditional resins because you have less surface, but you have a very fast flow which enables high productivity.

What our R and D team has done is the structure that you can see in the middle. So we kept the basic of a convective media. Basically, we can see that there's a few microns of diffusion, which enables to increase the capacity.

And the idea is to get the best of both worlds.

This can be highlighted here with the data I have shown in this graph. On the left, you can see the DBC and on the bottom, you can see the residence time.

So just to explain a bit more what I just shown, if you look at traditional convective media, like, for example, self domain for quenching, the flow rate has a very low impact on city, which you can see here. The BBC is more or less constant, even if you increase the residence time, because there's a very little diffusion part on the capacity.

If you take bids, it's very different picture, of course. So here you see here a strong impact of the residence time on capacity. And therefore, you have high capacity if you have sufficient residence time. So again, most of the time it's four minutes.

The idea of a convective membrane, which was developed internally, is really to shift this increase to the left of high capacity, short residence time. And that's exactly what we managed to do. You can see it here. The idea is to have a maximum capacity is roughly reach around one minute residence time. The actual residence time you want to use is more in the range of twelve seconds because you have already forty comparator, which is a relatively standard loading for the media.

Thanks to that, can conclude that at the end you will load a comparable amount of Mab, but you will go much, much faster because you have a residence time of twelve seconds against four minutes.

Thanks to this, you can actually do a lot more cycles than you would do with a resin. Typically with resins, you will do between three and six cycles to try to reduce the resin size because it's very expensive.

Each cycle is around two hours. Of course, have the setup around it. Some resins are faster, but it's good average.

If you go for RCC, for rapid checking chromatography on membrane, each cycle takes around ten minutes, maximum fifteen, depends a little bit on your feed stream. But at the end, you can do up to one hundred and fifty cycles per batch.

And that means you can read on size of a consumable. So of course, you have the benefits from a footprint perspective, but also very important from a perspective.

If you replace, let's say, a ten litre resin by a two hundred ml membrane, of course, it's much cheaper. So if you don't have the full use of its lifetime, especially for clinical trials, it's a huge benefit.

And that's actually the conclusion of that. The idea, as much as possible, is to use a full lifetime of the membrane within one batch that allows to have a very economical approach. Also, you have a shorter consumable, which means you have faster lead times.

And you don't need to bother with reuse cleaning validation. So it has a lot of benefits.

It also has a big benefit from a flexibility perspective, because at the end of the day, you can choose if you want to go for fifty cycles per batch and go fast. Then you can process everything within only one shift, sometimes even a few hours.

You can also, if you are so that approach typically is important if you transfer your process to a CMO, or if you have a production facility with a tight schedule and you want to make sure everything fits. So that can be a very good approach.

You have a more easy planning and you want to be very efficient on cost, you can increase the number of cycles to really fully utilize the membrane lifetime in one batch. And then it becomes extremely interesting from a perspective.

What you can also do is that if you validate the design space, let's say you are running late on your production because you had some issues in the steps before the capture and you have a few hours to catch up, membranes are delivered in cassette format, which is modular and very flexible. At the last minute, again, if you took care about that in your validation, you can add a couple of cassettes to win some time in your production and catch up with your schedule. So that's also a benefit.

I'd like also to introduce our multiuse rapid cycling chromatography system, which was developed with NovaSet at the time because it's a project that was three years old.

The idea was to use the BIO X platform basically to optimize the system to use membranes.

So you have different configuration, gradient or isocratic, depending on the number of cells you need. The idea is to really optimize all the volume, but also the valve switching time, which is very, very important when you go at high flows on small consumables to optimize your buffer consumption.

So the valves are reacting very fast, which is great for the membrane chromatography. Also, interesting thing, this modern system has UV spectrum sensor before and after the membrane, which allows to do a fingerprint of the batch quality, for example.

So that can tell you a lot about your process robustness. It can be very useful. Also, you can monitor for wavelengths if you know a given programmatic HCP, for example, absorbed at a given wavelength. So you have a bit of flexibility there. And very important, it's combined with SIM card nine. So it's a MBDA, so a multi variable data analysis software, which combines all the variable of the process variables within one.

And that allows to, after a few validation batches, you know what profile your batch is supposed to have.

That approach can allow to detect very early some process deviations.

So, for example, if you can detect early that the membrane has a problem because, for example, you didn't pre filter for that batch or any reason any reason there's a program. You can see it early and anticipate what you need to do. So that's a big benefit as well. And with that, I hand back to Ricardo, who will show you some performance data.

Thank you, Jeffrey. Exactly. In the next chapter, we look at some performance data for our newly developed ZartoBind Rapid A membrane.

So the first graph that I have for you here shows the dynamic binding capacity at ten percent breakthrough for different monoclonal antibodies or molecules which have an Fc region, so which will bind to the protein A ligand and the curves or it shows the dynamic binding capacity at ten percent at twelve seconds residence time for all those different molecules. You'll see here a range of different IgG1 monoclonal antibodies. We have an IgG2 in between, two IgG4s, IgE, and also two FG fusion molecules.

And what you see here is at twelve second residence time, we observed forty three gram per liter as an average and dynamic binding capacity with a minimum of thirty gram per liter, which has been tested so far and a maximum of fifty eight gram per liter for molecules.

So we could demonstrate here that we can achieve very good dynamic binding capacity, which is comparable to those of typical protein A resins. And also the observed variation, which you'll see here over the different monoclonals. This is also typical for protein A affinity media.

So then we have done also a recycling study with four different monoclonal antibodies using the same protocol which is displayed here for all those four antibodies. We started with an equilibration at the beginning using PBS and this was also used as a wash solution. Clarified supernatant has been used for the load and sodium acetate was used for the illusion where we collected eluate from one hundred milliAU to one hundred milliAU. And after each cycle, we cleaned or regenerated the membrane with zero point two molar sodium hydroxide.

You see here also the flow rates which have been applied for not so critical steps like equilibration and wash, use higher flow rates to really speed up the cycle. Here we use ten membrane volume per minute and the more critical steps like load and also elution have been done at five membrane volume per minute, which equals here for the load step to the twelve seconds residence time.

And you'll see also the different volumes which have been applied.

Of course, for the load that vary depending on the achieved dynamic binding capacity for the four different molecules and also the seed concentration which we had. Just want to mention as well that we have used an inline filter.

Here we used Zartopr two in front of the Zartoprint Rapid A nano device.

So in the next slide, you see the chromatograms or an overlay of the UV traces from the four different monoclonal antibodies. Here you see it displayed. We overlaid the two hundred cycles for monoclonal antibody one, two, three, and four. And what you can see here is that the cycles run very robust, so there is no shift in the UV traces.

And you can also see that the same protocol worked for all four different monoclonal antibodies and there was no, for the first stage, let's say no optimization implemented, which can still be done. But you see a very nice overlay, no shifts in the UV chromatograms. And this is the take home message from, this slide here.

On the next slide, we look into the trans device pressure traces for the four different monoclonal antibodies. And what you can see here that the maximum back pressure we received was the other monoclonal antibody III, where the maximum pressure was one point four bar or zero point one four megapascal. For the other three different antibodies, we only reached one point two bar back pressure. So that is very good. And the maximum pressure was reached with the ten membrane volume per minute flow rate. So even at the highest flow rate.

Yeah, and in general, we can say that there was no major increase in the pressure drop over the two hundred cycles. As you see that still also the transpressure traces here very well overlay for the two hundred cycles.

We have also measured the yield for the four different antibodies over the two hundred cycles, and this is displayed here. You see here, especially for MAP-one and MAP-two, they showed very high and very constant yield over the two hundred devices for monoclonal antibody IV, which is here the dark black dots. We achieved also a very good yield between eighty six percent, ninety percent, and only MAP-three showed a decrease here towards the end of the two hundred cycles in the yield. And this reason was because the monoclonal antibody was not stable for such a long time at room temperature, so it started to precipitate. And of course, this is later on, can be seen here in the yield which has been measured. But except for MAF3, we really achieved a very stable yield over the two hundred cycles. And even the third one was very stable until here between, let's say, around one hundred and forty cycles, it's really started to drop.

We have tested also the HCP removal for the four different feeds. You see here the load levels of the HCP for the four different monoclonal antibodies and here the lock reduction value, which has been calculated over the two hundred, yeah, every twenty cycles over the two hundred cycles. And here we can again see that for MAP-one and MAP-two, we reached a very high and also constant log reduction value between two and two point five log reductions. For MAP-three and MAP-four, it was a little bit lower, but still for MAP-four here close to two log reductions and for MAP-three around one point five log reductions. But you can also see that the levels of HCP in the feed was already much lower. So of course, the log reduction value won't be that high.

So also HCP removal was consistent over the two hundred cycles and is also very comparable to standard detected was the DNA removal. And here you see also a very consistent picture over the two hundred cycles for all four different monoclonal antibodies. So the HCP removal was, very good between two point five and three log reduction for all three different antibodies. Again, here you see the load levels for the four different antibodies.

And yeah, also the DNA removal is comparable to this of resins.

The last thing we have tested is the ligand leaching and this is displayed here in this chromatogram, in this diagram, sorry.

You see a very low ligand leaching with a maximum here around five parts per million. But in general, we can say it was very, very low. Our threshold was here at ten parts per million. So also ligand leaching of the protein A ligand from the membrane is very, very low.

So I would like to summarize all the findings here with this slide. You see here now displayed the different titers we had available for the four different monoclonal antibodies that was ranging between two point two gram per liter and four point three gram per liter. You see here the achieved dynamic binding capacity at ten percent breakthrough and at twelve seconds residence time. So that was also ranging between forty two gram per liter and even fifty four gram per liter.

The load level was chosen to be around seventy seven point five percent.

For this exercise here for this test, so we loaded between thirty two gram per liter and forty two gram per liter on the membrane. The yield summarized here, I mentioned earlier, MAP-three, there was precipitation occurring in the feed. Therefore, we give year two values, was very consistent for the first one hundred and thirty cycles.

So and then it was dropping because of the precipitation of the molecule for Map one and two very high values could be reached above ninety five percent. And Map four was a little bit lower in yield, but here, if we would do some optimization of the protocol, I'm pretty sure that we could also achieve here higher yields. You see the monomer content which has been detected as well, that was very high and consistent for all four different antibodies above or around ninety nine percent. HCP reduction you saw earlier, also htDNA reduction protein A is summarized here.

Protein A ligand leaching is summarized here on the slide as well. And I would highlight here the calculated productivity which we achieved that was ranging between one hundred and sixty gram per liter and two zero five to one hundred and seven gram per liter per hour for those four different antibodies. So very good and very high productivity could be achieved for those four monoclonal antibodies with Sartobind Rapid A.

We also did a comparison of the performance of our Sauterbind Rapid A membrane to a standard protein A resin. In this case here, we took one of the monoclonal antibodies which you have seen earlier. For this specific one, we achieved forty three gram per liter genomic binding capacity with our membrane. In this case here, thirty gram per liter for the standard resin and this was achieved at four minutes residence time compared to the four point three gram per liter which we achieved with the membrane at the twelve seconds or zero point two minutes residence time.

The yield was on a comparable level around ninety five, ninety six percent.

HCP reduction was also comparable for both protein A media.

DNA reduction in this case here was slightly higher for protein A, but still in a comparable level for both. Also, A ligand leaching was here in this case even lower for our membrane, but on a very low level for both different protein A media.

And what is really, yeah, the highlight here is that we saw that we see or get a forty point five fold increase in the productivity when you switch from a standard resin to CytoBind Rapid A.

And here I hand back over to Geoffrey.

Forgot to unmute myself.

Performance of course is very important, but it has no impact at all on your life if a solution is not scalable. So we are very happy and proud to announce that we have the first scalable membrane on the market ready to use in GMP format.

The portfolio from Software and Rapid A is based on the Software and Portfolio, which has been in the market for decades. Of course, constantly improved. We optimized several things since the launch in the late 90s, but nevertheless, we have a defined portfolio which is working very well on the market.

We have available ninety six well plates that are typically used for buffer optimization studies on automated robots. You can also have a PICO format of eight ml, often used for various clearance studies.

And then start to really, let's say, scalable portfolio. You usually develop a process on a nano.

The final nano is one point two ml. We first had prototypes of one. Now the final volume is one point two. And we scale up with capsules up to two hundred ml, which by the way, is enough to process standard two hundred liter bioreactor, which it's already to a decent scale. All this is commercially available. And we have a cassette as well for the large scale production of two thousand liters and more.

I guess it's, of course, stacked together and therefore scale to be the flexibility I highlighted before, where you can add one or remove one depending on last minute decisions if needed.

Very important to highlight is that the height of a scalable portfolio is constant.

That means the flow rate recommendations and the resulting residence time will be constant over our scale up, which is, let's say, the traditional way of scaling up chromatography. Also, important for the scale. It means the pressure of small scale is representative of a pressure you will see in large scale.

And that's, of course, important from a safety perspective at large scale.

You can see some data we collected here on the scale up of on the map.

So we did test on the nano one point two ml, the mini ml and on a five inch capsule. So a prototype of seventy ml. You can see the nano and mini are overlapping perfectly here because they are all under the same system, so same UV path and everything. So it makes sense for scalable products. The UV signal should be perfectly the same.

The five inch capsule was done on different systems. So the multiuse rapid cycling system I highlighted before, which is why it's not effectively the same because the UV cell is different. You cannot expect the exact same value.

Nevertheless, you can see it's comparable.

One small note is that it was a prototype of seventy ml. The last final shows are seventy five, which has a small impact on the final, but also on the illusion volume. You can see here the data. So of course we loaded the same amount, but that's obvious. Yield was constant, so ninety eight percent or ninety nine percent roughly. HCP and DNA reductions were also scalable.

The illusion volume was a little bit higher on the seventy ml device because, again, not seventy five here.

And also, the system was as a bigger hold of volume of a benchtop on that, but it's still comparable. And the productivity was concerned as well with one hundred and forty something gram per liter hour, which is a big improvement compared to traditional chromatography.

I'd like now to highlight some industrial considerations, because again, for us, it's extremely important to buy a solution that will solve bottlenecks at the industrial scale and not only in PD.

So the first thing I'd like to highlight is the hardware investment and footprint and the impact of back chromatography versus membrane. If you do bagged chromatography and assuming you have columns yourself, which, of course, can be avoided. That's limited to a given scale because you cannot get prepacked columns of one meter diameter, for example.

So if you pack up columns, of course, you need four tanks in a dedicated, often a dedicated room. You need a storage tank, obviously.

If you have manual columns, you need a storage packing station.

You often need also some accessories like something to move bed support, something to move the colon if it's a big one. There's a swing out device to move the colon tube for our providers, not from us, but that's also another device that if we, of course, you don't need to have it in production all the time, but nevertheless, it has to be there somewhere.

Of course, you have a current, but obvious and a chromatography system.

So usually one is used for production and HTTP, but sometimes you have one in the room of only five HTTP and one in production.

So that's a lot of hardware and a lot of footprint. If you go for a PDA, of course, you don't need to pack anything. So you just need a pilot order sets. Or if you go for tools at the two hundred liter scale, for example, you don't even need that.

And obviously you need a chromatography system for every single use, but multi use will do as well.

That's a lot less.

If you look at the footprint, so I took this from a a genetic location in BIO.

This shows the installation.

So we've a cut on with a skid, storage tank, cleaning vessel, car bags.

So that's a three d view here and on the floor up. You can see it here on the upper floor. There's also about the full tanks. So at the end of the day, this represents roughly one hundred and seventy five meter.

That was for a quite large column of one hundred and sixty centimeter diameter. But that's a big footprint.

That has an impact on your facility building cost, but also on all the maintenance and related costs.

So that's something you can get rid of with membrane chromatography.

Now, let's talk about the setup and handling as well. I don't know if everyone in the crowd has plan. I have and I can highlight quite well how it's happening. So if you pick a column, usually you take that opportunity to unpack the previous column. You do all the maintenance.

So you take out of your current of your production room to go to a packing room if there's one. Sometimes it's done in production directly.

The maintenance, of course, is often done by external suppliers, which means additional cost and planning constraints to your production schedule. That means, of course, you have to also pay for the spare parts of the print. You change mesh or center most of the time here and you check everything else.

After all the maintenance is performed, which often takes roughly one day and one day of unpacking, you have to install back the column in production.

That is done. So you connect everything. You do the safety.

The column needs to be clean. That's obvious, but also the system, the tanks. So again, that takes some time.

And the story needs to be cut. But to do that, you need to open a lot of bags of resin or drums, should I say.

Put them together, potentially change the buffer.

The buffer change is done between three and five times. It's a bit on the resin type, but it's a while because you mix your resin with a new buffer. You let it settle. You drain the buffer, and you repeat that operation three to five times. So again, a lot of efforts and manual operations.

Then you need to put on the system.

Resins are extremely sensitive to air, especially if you want to pass the HTTP.

So it happened on to me a couple of times that you have to flush the full system with water for literally more than thirty minutes, sometimes forty five, just because you still see some helpables coming out. So it's a bit frustrating because you're waiting to stop. So when that happens, you lose a lot of time and you consume lots of water as well.

I mean, it can be sometimes a challenge.

And obviously, you do a real operation here. That is a packing. So you will inject a story at a given concentration. You will reach a defined pressure or velocity depending on your method.

And then you will reach your final version factor. Once that's done, usually you often do a CIP after all. Sometimes you do directly with HTTP as a bit. But I there will be a CIP at some point.

So CIP takes, again, at least one hour and HTTP often the same.

Of course, then you may pass the HTTP and you're Okay.

But if you don't, which even happens to experience column packers. So some people have been packing columns for years, decades. And even then, it can happen that doesn't pass for any reason.

And then you start from scratch again.

Hopefully, passes.

And then you don't waste another three, four days. And you can go to operation.

So here, you will transfer back the colon. You will install it. So you connect all the pipes. Again, you need to do VR removal. So you need to be very careful with that. That's very important. And again, it takes some time and water.

A new CIP is often done before running the process. So liberation, load, wash, dilute, regeneration and few cycles most of the time.

Once all this is done, of course, you need to see IPR colon. And since you want to start, you need to rinse it with buffer and to storage solution.

So that's a lot of operations.

Can see there's a lot of things around the process itself. So it's not only the process that takes time. You should go for membrane.

So you can do a very quick installation because you just need to stack a few cassettes in a process holder. That's very fast. It takes maybe ten minutes. When you do a quick removal, membranes are not so sensitive to air as presence. So that means if you have a bubble, you can simply use the vent on the cassettes. So it's pretty straightforward.

If you like, you can do a quick CIP with low concentrated sodium hydroxide, for example,"