Continuous Chromatography of Peptides
(gentle music) - Hello, thank you for joining us. My name is Keith Porter. It's my pleasure to host you for the following Bachem webinar series, discussing the application and advantages of continuous chromatography in the purification of peptide APIs.
To start, I wanna first provide a quick introduction on Bachem. Bachem is a leader in the development and manufacture of GMP peptides and oligonucleotides. We are a company driven by innovation with heavy investments in infrastructure and cutting edge technologies. Bachem recently just celebrated their 50th anniversary this month. And with this 50 years of experience and expertise, we use it for providing products for research, clinical development and commercial applications.
Bachem is a global company with headquarters in Switzerland and locations in the United States, Europe and Asia. And we are on this Swiss exchange. So for more information about the company and investing, please check out our newly remodeled website at bachem.com. So it's my pleasure to introduce our guest speaker, or our speaker, Ralf Eisenhuth. He'll be joining me, I'm Keith Porter.
I'm moderator of this presentation. I'm a business development manager for the Atlantic territory in the United States. I live right outside of Charlotte, North Carolina, and I joined Bachem of March of this year.
So I'm new to Bachem, but I've been working with peptides as a synthetic chemist for the past decade. And again, I'm pleased to be joined by Ralf Eisenhuth. Ralf joined Bachem in 2008 as team leader in API manufacturing.
Since then, Ralf worked in various positions, including director of department large-scale purification at the Bubendorf site. In his current position, he is the process manager of chromatography and technology transfer. So we'd like to do a quick poll.
A quick questionnaire should pop up on your screen. Here we go. Sorry. Do you work on projects using peptides? So you should see a questionnaire pop up where you can select the answer and the results will be populated in real time and we should be able to present them very, very soon. So the choices are yes, no problem. Yes, but I experienced a lot of trouble.
No, but planning to; and no experience. So this is, do you work on projects using peptides? So it looks like a majority of everyone, 67%. Yes, no problem.
And then on the lower end 3% with no experience. So this area is new to you. So great. Welcome to this presentation and a couple more slides before I hand it over to Ralf.
All right, so Bachem's core strengths encompass projects relating to research and specialties. This is catalog business offering raw materials for solid phase peptide synthesis, as well as custom synthesis of research grade peptides and peptidomimetics. Our main focus is commercial API manufacturing. This is the development and manufacturing of new chemical entities, generic peptides, as well as small molecule active pharmaceutical ingredients.
And the last area is CMC development. So for those of you who aren't familiar, CMC stands for chemistry, manufacturing and controls. We offer a comprehensive range of services, ensuring regulatory and quality standards from the feasibility stage of projects through commercialization, are strictly adhered to throughout the life cycle of the API.
So our headquarters are in Bubendorf, Switzerland, and this is our largest peptide and oligonucleotide production facility, which is in Bubendorf. We also operate a large, small molecule API facility in Vionnaz, Switzerland. Both these Swiss facilities are GMP. We have GMP locations inside the United States, in Torrance, California and Vista, California.
And then we have a site in St. Helens, UK, which is between Liverpool and Manchester, which focuses on high-grade custom synthesis of research grade peptides. And we have business offices in Tokyo, Japan. At Bachem, we excel in innovation and innovation is the motor of our success.
So to continue to advance and to be the best at our industry, we consistently look to invest into growth strategies for our core business. Therefore, I'm pleased to introduce you to continuous chromatography of peptides and innovative downstream process, which Ralf will discuss. Before I hand it over to Ralf, I have one more slide. And it's just a simple bit of housekeeping.
So you'll see a button at the bottom of the screen, Q&A. For any questions that you have, just click on the Q&A button. Don't click on the chat button, click on the Q&A, and I will come back and facilitate these questions to Ralf at the end of the discussion. - Okay, now it's my turn.
My mouse is not so sensitive. And today I want to present some thoughts and data regarding the continuous chromatographic purification of peptides. After a brief introduction regarding pretty interesting trends in the peptide field, I will try to explain to you with the basic principles of HPLC separation and especially peptide separation.
Then we will go over to large scale purification of peptides, which is currently performed in batch mode. And after that, we will go to the interesting part of the presentation, MCSGP, where I will explain the process and show data from our initial case study that we did in 2017. After some additional considerations regarding the application of MCSGP in large scale manufacturing, we will finish with summary and conclusions.
So nearly two weeks ago, we had a big party here in Budendorf and in all other sites, because we actually celebrated the 50th birthday of the foundation of the Bachem by Peter Grogg. In the last decades, we specialized in medium and especially large scale manufacturing of synthetic peptides. And in 2019, we decided to start the manufacturing of oligonucleotides at the Budendorf site. In 2020, we manufactured the first API batch of an oligonucleotide with a batch size of one kilogram. We have over 150 NCE projects in the pipeline. So we notice early what is actually happening in the industry, especially when it comes to drug product volumes.
And that's really an interesting trend. In the last decades, the batch size for peptides was not that large, mainly due to the small daily dose, since they are quite active. So we had small volumes for our compounds ranging from three to something like 10 kilograms per year. In the last years, there was a quite interesting trend.
People started to actually achieve the holy grail of peptide drug application, meaning that now we have oral or even inhaled peptide drugs. And that's really increased the interest in peptides in general, and we see requests for processes that are generating several 100 kilograms to multi tons of peptide per year. This can be seen by the average size of the Budendorf facility, the main facility of Bachem, where you notice that more than doubling of the batch size in the last two years.
This is quite an obvious trend. And since we are now moving in the high volume fields, we are no longer flying under the radar regarding sustainability. Because if you have to manufacture something like three kilograms of an API, the process mass intensity is, it's just a blip in the waste stream of the overall industry.
Nevertheless, we are moving in the several 100 kilogram range. So now we see an increased focus on green chemistry and sustainability. Since we anticipated such a move, we actually started tracking the PMI of our processes some years ago. In the right side of the slide, you can see a typical solvent distribution for our preparative HPLC purification process of a peptide. In this case, it's Octreotide, a well known compound. As for all peptide cases, the solvent stream is mainly aqueous, meaning that of the overall solvent that we use more than 70% of water.
And the rest is modifier, which is just a little bit like 0.2% and organic. In this case, acetonitrile. Although it's mostly water, it's still waste, which has to be the... Sorry for that.
It's still a waste that you can't simply put down the drain. That's why we focus always on the solvent consumption of our processes. And here you can see what we achieved with an additional round of process optimization. In total, we were able to reduce process mass intensity in kilogram, per kilogram of trucks, substance by more than 40% for the Octreotide process. Imagine that you want something like a ton of Octreotide.
This is more than one million liters of waste that was saved due to the process optimization. In addition, we are always investing in capacity to fulfill the increasing demand. And for this I plotted you the increase in large-scale purification capacity at the Budendorf main facility of Bachem.
And here, you see I'm taking 2017 as baseline that we more than doubled purification capacity until now, 2021. This was mainly achieved by increasing batch size, column size, but also by building new purification lines. Since most peptides are still isolated by freeze drying, we also had to double the freeze drier capacity at our Budendorf facility, going from 100% as a baseline in 2017 to 222% in 2021. All this was achieved in the current large scale manufacturing building in Budendorf, which is highlighted here.
It's a TIDES facility with 140,000 square feet, but space, it's getting tight there now. That's why we actually start building a new manufacturing building here. It's the so-called building chain. Came Budendorf and the construction will start next month. And it will be operational in 2023. We will have another 170,000 square feet for TIDES manufacturing, meaning peptides and oligonucleotides.
In addition, we will have additional administrative buildings and something that I never thought would happen when I joined Bachem more than 10 years ago. We actually have campus parking now. Overall, we will invest something. We will invest more than 400 million Swiss francs globally in capacity increase. So, now that you've seen that we will manufacture more and more peptides, we have to think about how we can do this in the most efficient way.
Before we start with MCSGP, I want to explain you a little bit about the basics of HPLC separation, because this is always very interesting. In the beginning, we have two compounds marked in green and red. This a mixture that is put onto a column, and then we see a typical chromatogram where the peaks are hopefully separated. There are two factors that are mainly known to analytical chemists, but they are of prime importance for preparative HPLC purification. That is a retention factor, which is calculated with the formula for kx. And thus selectivity, which is simply defined by the retention factor of the later eluting compounds divided by the retention factor of the early eluting compound.
In general, selectivity simply describes how the column can discriminate between two substances and selectively retain them. By definition, it has to be larger than one because if it's smaller than one, you will not have any separation. It is the goal of every purification development that you achieve maximum selectivity for the target compound and its related substances. For analytical purposes, there's something like the holy trinity of selectivity, which is defined by the interaction between this structure, the stationary phase and the mobile phase. Regarding the structure, peptides are quite interesting because there are a lot of options to actually achieve selectivity. You have hydrophobic parts in the peptide.
We have residues that can induce pi-pi interactions. We have H-bond donors, H-bond acceptors. We have compounds that can interact with anions depending on the pH. And we have compounds that can interact with cations depending on the pH. In addition, we have a dipole moment in the molecule and we have specific molecular shapes for the target compound and its related substances.
So, these are a lot of options for selectivity. And luckily, the stationary phase, as long as it's silica based, also has quite a lot of options for selectivity. We have aromatic selectivity. We have CH2 selectivity, meaning how good the column can discriminate between a compound that has one CH2 group mob, and we have shape selectivity, hydrogen bonding and ion exchange capacity. These are results from the so-called Tanaka tests that we do in-house for column characterization.
In general, you can split the selectivity of the stationary phase that is induced by the surface modification and the part that is induced by the silica. In this picture, you will see the Tanaka results for two C18 columns, and they look pretty similar actually. That's why we always try to screen different and surface modifications to actually have columns that are as different as possible. In this case, these are the results for phenyl and a C18 column.
And you see the phenyl column is in yellow, as it has suddenly aromatic selectivity that is not present on a C18 column. And here you see that those columns are really different. And that's why you should always focus on the surface modification for screening and not on the manufacturing process of the supplier. It happens quite a lot, but people tend to screen something like six C18 phases and really try to get the best purification out of columns that are not really that different.
So for the mobile phase, which is probably the most interesting part of the holy trinity, we have also an inherent holy trinity of the... For selectivity, we have the counter ion, which will define the hydrophobicity of the molecule, which is the main interaction for peptide purification. In addition, the counter ion can block interactions between the molecules and the stationary phase. Then we have the pH, with which you can change the number of charges on your molecule. And it will also determine whether the silica surface will actually interact with your molecule. And then there's organic modifier that actually modulates the strength of interaction between the stationary phase and the target molecule.
In addition, this can be used to suppress unwanted interactions as the analytical chemists in this chat will probably know. If you use acetonitrile, you will suppress most pi-pi interactions. So if you have a closer look at the holy trinity, we have a perfect match between the stationary phase selectivities and the peptide itself, which is modulated by the mobile phase. And there you can have the options for the organic modifier, the salt modifier, the ionic strength, which is also very important, and the pH. Based on this consideration, it's clear that the same effort for mobile face screening is done as far as stationary phase screening.
Well, now that we think that we know what is actually happening, let's have a look at a real purification. This is Octreotide purification, where we go from 87% purity to 99.5% purity. And in this specific case, we actually do a quite a lot during the purification process.
We form a disulfide, which hopefully explains the change in retention time for the compound. We also convert isopeptides during this step. And it's a one dimensional purification with impurity limit of smaller than 0.1%, which is basically small molecule specifications.
This is a clear trend for smaller peptides up to, I would say 10 to 14 amino acids, where more and more small molecule specification, at least for generics come into play. Well, the bread and butter of purification is still batch mode purification. In this slide, you see one of our largest column. That's a 60 centimeter column. And the purification actually happens in the metallic steel cylinder, which has a batter age between 25 and 30 centimeters. So to have an idea what is actually happening, I've drawn this quite sophisticated machine.
Oh, sorry. That one, it's one. What is actually happening, we have a crude peptide solution that is put into the cylinder and then we have created, and then we have hopefully the separation of the impurities marked in other colors from the target compound in green.
To understand what is happening in the column, I actually draw quite a simplistic version of the column. So it's this one, this gram. And here, I want to explain you what the process steps that are part of a preparative HPLC purification. You should remember most of them because when we discuss MCSGP, you will see each of those process steps happening also during MCSGP.
So the beginning of every one, that's actually the column is actually operated with eluent containing not a lot of displacers so that the peptide gets absorbed onto the column. And once the peptide is absorbed into the column, you start the gradient or the isocratic illusion, and the initial band actually gets split up into five zones, starting with a black zone waste, where no product is in, then we have a zone that is called weak, then we have a zone that is called target. And one that is strong and one that is waste. At the end of the column, we hopefully have distinct several bands and we discard the black bonds and we collect the blue, green and red ones in three different vessels.
After that we do the cleaning in place. The blue ones are called weak side cuts and the red ones are called strong side cuts. A side cut is actually a fraction of the peak that contains the target compound. But the fraction itself does not meet the criteria for further processing, meaning either a second HPLC purification of at least trying to isolate the drug substance. Nevertheless, this fraction still contains enough target compound to actually allow purification.
In the peptide field, this is a very common practice to get a lower cost of goods. Nevertheless, the side cut fractions contain related substances that are difficult to remove from the target compounds, because they are structurally very similar. Thus in most of the cases, the product generated from side cut repurifications tend to have a lower purity. In the graph, I've collected the data from purification campaigns on one of the large columns.
The green dots are batch purifications, and here, you can see the yield purity trade off, meaning that the further you go to a higher yield, the lower the purity of the compounds will be. And in this case, we did a full campaign, including side cuts purification, and we were actually able to increase the yield by more than 5%. But the trade off was that we actually had to be repurify 2.5 cubic meters of side cuts. So this is something that is often overlooked when it comes to scaling of preparative HPLC processes, because the amount of site cut fractions will scale linearly, and this will be a challenge when it comes to very large scale purification campaigns since you need a lot of work in freezer space. That's why we thought that MCSGP, actually with its inline recycling of side cuts fractions is a very interesting solution for this very specific scale-up challenge, MCSGP is a creation for multi column countercurrent solvent creating purification.
So here you see one of the small developmental systems from the ChromaCon age here. It consists of several pumps to the... At the bottom there is actually the gradient pump. And you see that it has two columns, and this is very important because if you have two columns, you can actually do quite a lot of stuff in parallel.
We were quite interested in this technology. And in 2017, we rented a cube 30 from the Chroma Carnegie and did an initial proof of principle study. As a target compound, we chose tetracosactide, which is a legacy process that has very challenging impurity profile, and the throughput of a single shift team on one of our largest columns, meaning 60 centimeters in the lower single digit kilogram range per week. That is not a lot. The purification process itself is quite complex.
It has two orthogonal HPLC dimensions and one final solid extraction step. In the first dimension, we actually have to apply something that is called an inverse gradient to achieve the target purity with acceptable throughput. So how does MCSGP work? I will now show you an animation for something that is called steady-state.
I will explain to you what this is a little bit later. So basically we have two columns. So we can do a lot in parallel. In column one, we have the crude and the weak and strong side cuts absorbed onto the stationary phase. And in column two, we're actually doing the cleaning in place that is usually performed after correction of the fractions and the target compound. So while column one separates the bands that I've shown you earlier, column two gets the cleaning in places performed and the actual operation.
Once the waste actually reaches the end of the column, it gets discarded, and when the fractions that are designated as weak side cuts get to the bottom of the column, the columns are put into series and the material over the weak side cuts gets transferred onto column one. During this step, the solution has to be diluted with eluent A to ensure reabsorption of the peptide onto the stationary phase. While the target compound is collected and the column two is fed, and afterwards the strong side cuts get put onto column two and also diluted with eluent A. So at the end of this sequence, we're actually in a position where column two has the same composition as column one in the beginning, and column one has the same composition as column two in the end.
This is called a switch. And you can imagine that if you do the same on column one, you end up with the same conditions. If you do the same with column two, you end up at the same conditions on column one, and this is called a cycle. So, what we did when we had the system, we converted the first dimension of the purification process using the built in system software. Since I had a lot of experience with the separation, I've chosen to use a lower load to have an easier separation.
But the lower load was actually a more than overcompensated by the automation of the system itself. After four injections, we reached the steady state of the purification, meaning that your profiles look absolutely identical. A steady state in this case means that the same quality goes in and the same quality goes out. This is one of the main advantages of continuous chromatography, since once you reach a steady state, you are actually set regarding the reaching of the specification. In total, we did 45 injections since we were interested, rather the process is stable and the data that we generated looked very convincing. To the left, you see the purity as determined by HPLC and after the ramp up, which is actually the first illusion, we are in a range where nothing changed as shown in the regression.
As well, there's no trend in this. And also the concentration remained largely unchanged. And this is a very good indicator that your process is stable. In this specific case, we were actually able to surpass the target purity of 97.5%, or already after the first dimension.
So we were able to omit the second dimension, which of course saved a lot of solvents. And the purity was higher compared to the batch process, and this yield was comparable. This is a very good result for not optimized process. And you can imagine what could actually happen if we would have put more effort in this process, we could actually end up here, meaning that we have higher yield at comparable purity, and this is very good for the cost of goods.
The solvent consumption was reduced by overall 33%. This is also very good for proof of principle experiment. And mainly we were able to reduce the water consumption. There are also some very important KPIs that are often overlooked when it comes to preparative HPLC purification. And one of them is what I call the total volume to be handled for further reprocessing, meaning these are main cut fractions and side cut fractions, and for tetracosactide, I calculated it and the area for one kilogram of drug substance, we actually have to store more than 1000 liter of main cuts and side cuts for at least one day.
And there, we were able to reduce this volume by nearly 90%, which is incredible. In addition, we were able to increase the product concentration by a factor of three, meaning that we actually de-bottlenecked another critical step, meaning the freeze drying. And as to be accepted for an MCSGP process, we reduced the volume of side cuts by 100%.
But based on this quite amazing data, we actually bought the first cube system in 2018. This is the one to the left, where the set up four or five 20 meter columns on this system, we manufacture non-GMP scale-up batches, and the largest batch we manufactured so far were 300 grams of API. In 2020, we bought another system. This is the cube 30, and the one to the left is the cube 100. And this system is also used for process development and scale-up studies.
In 2020, we also ordered the first large scale GMP systems for oligo and peptide purification. And they are planned to be operative at the end of Q4 2021, meaning in less than six months. Nevertheless, before you start ordering cubes, you have to consider some limitations, at least what I think. If you want to apply MCSGP, you will have more upfront investment compared to the two additional batch mode. You have to investigate the stability of the feed solution, especially with the formation of aggregates or precipitation of some fairly solubilized related substances. And of course you have to know if critical impurities, mostly stability indicating compounds are generated during the storage of the feed or the storage of the fractions.
And contrary to batch purification, you also have to establish a column lifecycle management program very early. If you don't do it, you may end up with something like this. In this case, I plotted the recovery so far, meaning that how much of the peak we actually... Of the product that we injected into the column we actually recovered, and there's a clear trend to lower recovery. And in this specific case, the column actually deteriorated due to a precipitate.
So if I may summarize this talk, there's a clear trend to delivery forms that are not based on syringes. And this trend will lead to large batch sizes in the peptide field. We always had a pressure on cost of goods, but there's also now environmental considerations that are becoming the major driving forces in process development. If we want to have sustainable purification processes, meaning that we apply something else than acetonitrile for the purification, we will need a disruptive technologies because the separation will be poor. We had a close collaboration with YMC and ChromaCon over the last several years. And we think that MCSGP is such a disruptive technology and that it's ready to apply it at scale for peptide purification.
At the end of 2021, we actually plan to have the systems that remained mentioned in the press release by YMC active and ready for GMP manufacturing. With this, I'm done with the presentation. I thank you for your attention and back to you, Keith. - Perfect. Thank you, Ralf. I see, we have a couple of questions, but I think there's another poll.
I'm gonna use my keyboard. Let's see. Yes, there is. Okay. Is upscale of peptide API manufacturing a topic for you? So again, you will see the question pop up on your screen and we'll get real time results after 10 to 20 seconds. So is upscale of peptide API manufacturing a topic for you? And the results, so, yes.
So, well, 50%. It's exactly what you're looking for or that you're very interested in. So 95% total, yes. 5%, no. And then 0% need more information. So thank you for the time, everyone that contributed to the polls and thank you to everyone who's asked questions.
So I'll pop some of those up right now. So if we don't get to your question, you can always email us at firstname.lastname@example.org. But the first question, will MCSGP be always the best solution for any purification process? - This is an interesting question. I think anything but chromatography is still a topic in the industry. The first thing that should actually be evaluated is whether you can precipitate your peptide and purify it in that way.
If you have not a large volume product, it may not be worthwhile to actually apply MCSGP for the purification. And also if you have processes where you have very good recovery, during the batch purification there, you may be better off to automize the batch purification process. - What do you think are realistic increases in yield compared to developed batch processes? - Yeah, this is interesting. It really depends on the quality of the developed batch process. We pride ourselves that we have very good batch process.
So now the increase in yields is most likely 5% more than what we would have achieved with the batch process and recycling of side cut fractions. - This one specifically towards our use of it at Bachem. You showed the increase in, it's a long one. You showed the increase in column capacity over the last years at Bachem.
Do you see the need for an even larger column to satisfy the need of larger batches? Will MCSGP be a viable solution with a much smaller footprint? - Well, currently, I don't think that we will move to something like two or even 120 centimeter columns. Because if you have such a large column, you actually need a project that needs multi tons per year. And I personally think that MCSGP is the better answer to the increased demand. I don't think that we will move to two meter columns. It will most likely be some sort of continuous process. - Okay.
And will the use of MCSGP increase or decrease the footprint of your process? - It will decrease the footprint of the process since it's not only the column and the chromatography skit that actually defines the footprint of a process. You also have IPC lab that has to analyze all the fractions that you generate, you need walk-in freezers. So MCSGP will definitely decrease the footprint of the process, compared to a traditional batch chromatography. - All right, this attendee has a follow-up.
Does it require special, solid face packing material or can it generally use available packing material? - You can use whatever you use for batch chromatography. - Perfect. Well, I think that's it for time. We've answered five or six questions and there are a couple more, but I think if your question wasn't answered, please feel free to email at email@example.com. And I thank you and I'm sure Ralf thanks you for your time.
And check out our website, our newly remodeled website, www.bachem.com. So thank you everyone for your time and have a good evening or a good day. (gentle music)