As part of BioProcess International Digital Week, Astrea Bioseparations CEO Steven Burton presented “The Multi-Mode Mimetic Ligand™ Library: A New Tool for Rapid Development of Downstream Processes”.
Lauren Berry (BPI host): Hello everyone and welcome to BPI Digital Week, brought to you by the producers of the face-to-face BioProcessing International conferences, visiting Tokyo, Santa Clara, Vienna and Boston next year. My name is Lauren Berry and I’ll be your host for today’s session titled: “The Multi-Mode Mimetic Ligand™ Library: A New Tool for Rapid Development of Downstream Processes”.
First, I’ll cover some quick house-keeping items, if you experience difficulties with audio or advancing slides refresh your screen with F5. If you are experiencing other issues, hit the question mark button to receive assistance. At any time during the presentation, submit your questions into the Q&A window on the left-hand side of your screen. In 24 hours, you’ll receive a link to watch the recording of this session, you can also download the presentation slides in the resource list box to the right of your screen.
Let’s now begin by introducing our speaker, Dr. Steve Burton, CEO of Astrea Bioseparations. Thank you for joining us today Steve, now I’ll hand it over to you to begin the presentation.
Steve Burton: Thank you Lauren, and hello everybody. Today I would like to present a new tool, The Multi-Mode Mimetic Ligand™ Library, which enables rapid and inexpensive identification of chromatography adsorbents for use in downstream process developments.
Before we go into the presentation, I’d first of all like to say a few words about Prometic, some of you might not be so familiar with our company.
So, Prometic is a Canadian corporation with its head quarters in Montreal in Canada and the company has a focus on plasma protein therapeutics and small molecule therapeutics, and it also has two operating subsidiaries which are semi-autonomous subsidiaries, one of which (Prometic Plasma Resources) is focused on plasma collection for use in our plasma fractionation programs and the other is Prometic Bioseparations which is focused on the development and supply of products and services for use in biomanufacturing. So today we will be focussing on some of the work that we’ve been engaged in as a part of the Prometic Bioseparations group.
So, Prometic Bioseparations or PBL is UK based, we have our research and development facilities based in Cambridge in the UK and we have our GMP standard manufacturing facilities located just off the north coast of the UK on the Isle of Man. We have been developing and producing chromatography adsorbents now for more than 30 years and we focus on the high-performance chromatography adsorbents, packed columns and the development of efficient and cost-effective downstream processes. Currently in Prometic as a whole we have around 500 employees and we have around 100 employees in the UK as a part of the Bioseparations group.
So, we produce a number of products ranging from small screening column blocks, and I’ll be saying much more about that over the next few minutes. We also supply smaller amounts of material, either in packed columns or slurry packs for use in process development. We of course, supply large amounts in bulk quantities for use in large scale bioprocessing and recently we have launched our own range of pre-packed GMP ready chromatography columns called EvolveD™ for use in continuous processing and continuous chromatography.
The company has adopted a GMP approach to adsorbent manufacture, we have very high standard production facilities, all of our production areas are ISO8 standard or better, we are ISO:9001 registered and we work in conformance with the ICHQ7 guidelines for the manufacture of active pharmaceutical ingredients, so that’s a very high working standard for the production of chromatography materials, and that reflects our commitment to provide high quality materials for use in bioprocessing.
Currently we have the ability to produce batches of adsorbent from 1L all the way up to 1,000L, in the photo here you can see one of our new 2,500L reactors that’s been installed at the Isle of Man plant and that gives us a manufacturing capacity for chromatography adsorbents in excess of 35,000L per annum, so we are very well placed to be able to serve the needs of the industry and currently over 18 of our products are now used in relation to regulatory approved products and processes, both FDA approved and EMA approved, so we’re very familiar with the quality standards required for these materials and their applications in biomanufacturing. We have some products that are used now in chromatography columns up to 2m in diameter, so again very well experienced in the use of these in production of biopharmaceuticals.
I mentioned just now the EvolveD™ column, this is a new product range for Prometic, these columns are pre-packed, GMP ready, ready-to-use columns. So they enable a plug and play approach to column chromatography, they come fully qualified and they are very simple to use, at the end of their life they can be then discarded and replaced. So that represents significant time and cost savings to users, particularly to companies that use large numbers of columns or use relatively small columns which are replaced relatively frequently. These columns are available in a range of sizes, currently we have column diameters of 7, 10 and 20cm and we are in the process of developing a 25cm diameter column at the moment. If you’d like to know more information on those then please contact Prometic and we would be very happy to help you.
Now moving on to the main body of the presentation, I’m sure many of you are very familiar with bioprocessing and you’ll be aware of the various techniques that have been used to produce high purity proteins. These techniques vary in selectivity, one end of the scale we have the filtration type techniques, originally gel filtration but these days it’s ultrafiltration or diafiltration that allows molecules to be separated on the basis of their size. We then have, moving to a slightly higher level of selectivity, we have techniques such as ion exchange chromatography or hydrophobic interaction chromatography which enable the separation of proteins on the basis of their net charge or their net hydrophobicity. In more recent times, there’s been a move towards the introduction of other sorts of ligands that go by a whole variety of names but their commonly referred to as multi-mode ligands and these are small, simple ligands that generally combine charge and hydrophobic properties and these are able to interact and differentiate proteins on the basis of their surface charge and surface hydrophobicity and the special relationship between those charge and hydrophobic groups. Then finally we have what currently represents the ultimate in protein purification, which is affinity chromatography, where we have affinity ligands that are able to bind selectively to specific sites on a protein and that selective separation enables very high levels of purification in a single unit operation.
I’d just like to focus more now on the multi-mode ligands and the way in which these have been used in downstream processes. Not too long ago downstream processes were generally represented by a series of steps, each step provided a progressive increase in purity, so it wasn’t uncommon to use, starting with some kind of precipitation technique then ion exchange, HIC, size exclusion, maybe affinity right at the end of the process and eventually having a protein with the required purity. But that isn’t a very efficient way of producing proteins because the losses of these types of processes tend to be quite high. These days we use a slightly different approach, we use generally an affinity capture step right at the beginning of the process to give us a very high level of purification in one step. Then we follow that up with 1 or 2 polishing steps to remove residual impurities and those polishing steps can be ion exchange steps or HIC steps, but increasingly these polishing steps are represented by multi-mode ligands and they’re finding a niche now in this role, not so much as in a capture step but certainly as a relatively high resolution way of removing residual contaminants in an efficient way.
I could spend a lot of time talking about affinity ligands, Prometic has been working in this area for 30 years plus and we have huge experience. I just wanted to share 1 slide with you just to show how powerful affinity chromatography can be, you’re probably all familiar with the use of protein A for the capture and purification of monoclonal antibodies. This slide just demonstrates that we can use the same affinity capture approach with a different adsorbent for other proteins. So, in this case it’s a Plasma protein, and it’s a protein called Plasminogen and we were able to capture very high efficiency Plasminogen using a designed affinity ligand that we developed. This enabled a 340-fold purification of Plasminogen directly from Plasma in one step and that gave us a purity of well over 90% and a recovery of well over 90%. So, you can see there, that with the right type of affinity ligand we can develop very effective affinity capture steps for a whole variety of proteins, not just monoclonal antibodies.
So, thinking more about the polishing steps now and the use of multi-mode ligands, these are small molecule ligands that can participate in two or more types of binding interaction and usually they’re small ligands that combine a charge group, either anion or cation in close proximity to a hydrophobic group, which is usually some kind of aromatic group. These multi-mode ligands can be used as a way of enhancing the selectivity of ion exchangers and some of you might be familiar that by adding this hydrophobic group, it enables ion exchangers to be used at a higher ionic strengths in the presence of higher concentration of salt for example. They can also be used for binding and elution of proteins, that requires adjustment of the ionic strength by additional removal of salt or the adjustment of polarity by adding or removing polarity reducing agents such as solvents or detergents, and also pH adjustments and it’s not uncommon with multi-mode ligands to have to spend quite a significant amount of time with the design of experiment type approach to fully optimise and identify the working window for binding and elution for these sorts of ligands and that is required because generally speaking there are relatively few multi-mode ligands available. The multi-mode ligand shown here is actually the GE Capto MMC ligand that some of you might be familiar with, and certainly to apply that one ligand to the purification of any given protein is going to require fairly extensive optimisation of the binding and elution conditions.
So we were looking at this situation and were aware that we have amongst our ligand libraries, many thousands of ligands with very different types of protein binding groups and many of these can be used in a multi-mode fashion and so that led to the that well, rather than choosing a single ligand and adjusting the conditions to suit that ligand, why not take the protein solution that we have that we need to purify and find a ligand, a multi-mode ligand that is suitable, that is optimal, for that particular protein in that particular mixture, so it’s a very different way of looking at multi-mode ligand chromatography.
So what we have done here now with our Mulit-Mode Mimetic Ligand™ Library is we had a much larger array of ligands, I’ll show you in a moment our multi-column library block and we have 96 different multi-mode ligands in that block and these ligands are typically attached to agarose beads by the use of spacer arms to make the ligands fully available for protein binding. We use a triazine attachment chemistry that gives very high-fidelity, so it gives very good synthesis and purity of the synthesized ligand and we use a solid phase synthesis approach to produce these ligands and that gives us, this chemistry gives us alkali stable ligands, so all of these ligands can be treated with 1 molar sodium hydroxide and the majority could even be autoclave if necessary a neutral pH. We use a wide variety of groups, so we’re not just restricted to one or two anionic or hydrophobic groups, we can use a whole range of different groups here, anions, cations, hydrophobic groups, hydrophilic groups or any combination of these.
The libraries themselves are packed into 96 column blocks, you’ll see some photos of this in a moment, and these blocks are micro-column, so each one is 250 microlitres, so that’s 96 discrete columns in a single block which is in the standard 8 by 12 array that is commonly used for screening and for assays. They are true packed columns, so they have a top frit and a bottom frit and they can be run in a chromatographic flow-through mode so these aren’t batch adsorption plates, these are true miniature chromatography columns that can be operated in a true flow-through mode. Screening is very simple, we can do this by running the column blocks under gravity and the target feed material can really be anything, we can screen either ligands for capture or for binding impurities and it’s probably the use of these ligands for polishing applications that is the most interesting right now because for those application we would simply want to bind the impurities and we don’t want to bind the target, so as long as the target protein passes through the column unhindered and the ligand binds impurities then that’s a potential candidate for use as a polishing step.
These multi-mode ligands now are a little more selective than the traditional multi-mode ligands and they sit here in this array of different purification techniques, so they’re not quite as selective as the affinity type ligands but they’re more selective than the traditional multi-mode ligands. In terms of synthesis, I mentioned the triazine chemistry, so we can see that here, we start with agarose beads which are very hydrophilic, very robust, very stable, and then we attach to those beads a chemical activating group which is sodium chloride and that gives us this dichlorotriazine functionalised material as you can see here. Then we can attach to those reacting fluorines different amine groups and these amine groups, the side chains of those amines can really be anything, they could be an ion exchanger type compound, it could be hydrophobic or even other types of groups that we use in our affinity ligand library screening applications. When we react the first amine compound the reactivity of the remaining chlorine is a lot less than the reactivity of that first chlorine, so that means that we can drive these reactions to completion so we get a very clean synthesis, the attachment of the first amine we wash away the impurity and the unreactive materials and then we can react the second amine with the remaining chlorine to give it this dye substituted ligand product.
The types of our groups that we have with these ligands are really quite a wide range now, so in terms of ionic groups we can have sulfonates, carboxylates, phenolate, some amino, pyridyl, various types of functionalised amino groups here and in terms of hydrophobic groups we can have phenyl groups, we can have alkyl groups, and we can have other types of groups as well, such as cyclohexyl groups. By combining these two groups of these ionic and hydrophobic side chains we can produce a library of multi-mode ligands that have quite diverse structures and quite diverse protein binding properties.
We’ve organised these ligands into six different zones, and the reason for doing this is to make it easier to understand the chemical structures and chemical properties of the ligands in each zone and how you might approach the binding and elution of proteins from those zones based upon knowledge of the types of ligands that are in those areas. In zone 1 we have ligands that are predominantly ion-exchange in nature. In zone 6 we have ligands that are predominantly hydrophobic in nature and then in zones 2, 3, 4 and 5 we have ligands that combine combinations of ionic groups and hydrophobic groups. So in zone 2 we have ligands that combine anion and anionic groups with aliphatic groups. In zone 4 we have ligands that combine anionic groups with aromatic groups. In zone 3 we have ligands that combine cationic groups with aliphatic groups and in zone 5 we have ligands that combine cationic groups with aromatic groups. So we have quite a diverse array of ligand structures and different ligand binding properties.
We did some work to convince ourselves that these ligands really did have different ligand binding properties and the way we did this was to analyse the chemical structures of the ligand using some computational chemistry software, the schrodinger suite, and this software is very powerful and it enables modelling of ligand structures, and protein structures for that matter, and it also enables the characterisation of ligands on the basis of a whole range of different chemical descriptors, so we can look for hydrophobic accessible areas on the ligand and that’s shown there in the top left, blue means it has a very low score for hydrophobicity, red means it has a very high score for hydrophobicity, and you can see that within those 96 ligands we have a very diverse range of ligands in terms of the hydrophobic properties. We can do the same for other chemical descriptors such as hydrophilic accessible surfaces, you can see we get a completely different there for hydrophilic ligands. We can look at ligands that can participate in hydrogen bonding, in terms of hydrogen bond donating groups and ligands that participate in hydrogen bond accepting groups and again we see very different patterns, so this is telling us that these ligands are different, so when we challenge them with different proteins they’re not all going to behave the same, some will bind certain proteins and some will bind other proteins.
We can look at more and more of these types of descriptors, Pi carbons is a measure of the aromaticity, so as you would expect the groups that have predominantly aromatic groups are lighting up there in red. We can look at overall polarity, log P, this shows us the ones that are mainly hydrophobic. We can look at the number of rotatable bonds in the ligand that gives us a measure of flexibility of the ligand. We can also look at predicted binding to different proteins such as Albumin and if you look here you can see three of these patterns are actually quite similar, log P, the number of rotatable bonds and predicted binding to Albumin all look fairly similar and you can see there’s one ligand in the C11 position that’s lighting up very strongly in all cases and this is because this particular ligand has very long hydrocarbon chain for the aliphatic group and if we think about Albumin, Albumin is a protein that binds long-chain fatty acids which are long-chain hydrocarbons, so what we see here is that this particular ligand at the C11 position we’re predicting it’s a good Albumin binder because it is relatively apolar, it has a long aliphatic chain and that makes it quite flexible as well. So now there’s chemical analysis by using computational chemistry it can be very powerful to predict how ligands might interact with live proteins and what to expect when we do screening, so we’ll come back to this Albumin prediction in a little while.
This is what the Multi-Mode Ligand™ Library looks like, as you can see it’s in a conventional 12 by 8 array with the various zones colour-coded there. One of the main things to emphasise about this is it’s very quick and easy to use and it’s not necessary to use a robotic liquid dispenser if you have one you can use it but it’s necessary, you can just use the multi-channel pipette as is shown here. The blocks are running the gravity so you don’t need to use a plate centrifruge to get flow through the beds here, they will just drain all by themselves under gravity and if you have plate based assays already set up for your proteins then it’s possible to do the binding and elution studies very quickly indeed, I mean that we typically do these studies in two or three hours in our lab and if you compare that to how long it would normally take to run 96 columns you can see that’s a fantastic saving in time. As I said these are discrete chromatography columns so they have a top frit, they have a bottom frit, so once the seals are removed the plates will then drain quite easily under gravity and then it’s just a question of equilibrating the columns and adding the desired feedstock and just letting that flow through the columns under gravity into a collection plate, and we provide a collection plate with the column blocks as part of the kit.
The starting material should be the actual fraction that you want to purify so that could be your cell culture supernatant, it could be fermenter broth, it could be Plasma or serum or whatever it is, it could be an existing column fraction you have if it’s a polishing step that we’re developing. The key thing here is that we start with what we want to purify, not some sort of artificial mixture which contains a purified protein in buffer, that doesn’t give generally speaking, a very good prediction of which ligand that which adsorbent would be useful for purification because what we want to do, as well as looking at the binding amount of target protein, we need to understand how other compounds, other impurities that are present bind as well, so if we want a polishing step we want to know that the target protein doesn’t bind and that the impurities do bind. Conversely, if we want a capture step we want to know that our target protein does bind and that the impurities don’t bind, and the best way of determining this is by using the starting feedstock that has both the target protein and the impurities present.
In the first example of the use of the Multi-Mode Mimetic Ligand™ Library, we’re going to look at screening of a monoclonal antibody feedstock, so this is a CHO human IgG1 antibody, it’s in clarified cell culture supernatant. We start by equilibrating all of the columns in a phosphate buffer at pH 7.4, we then load our feedstock onto this, so we’re loading 1ml. The columns are 250 microlitres, the wells are 1ml in volume, so we can load that as one 1ml fraction, if we wanted to we could load more than that, just simply by after that first ml has drained through the columns we can then pipette on another 1ml fraction. So you can do that as many times as you want to increase the loading volume if that’s required.
After loading, we then wash the columns, usually that washing is done with the equilibration buffer and then finally we will then elute off the bound proteins. In this case we are using an acidic buffer for elution, generally speaking that’s because acidic buffers work quite well for antibody elution off a whole range of different types of adsorbents, particularly Protein A and so we used an acidic elution step here. We can use other sorts of elution buffers depending upon the nature of our protein, we have here with these multi-mode ligands because they provide embodied combinations of charge groups and hydrophobic groups, very often an elution buffer that combines both salts and a polarity reducing agent works quite well and I’ll give you some examples of this later on, but in this particular example we’re eluting with a citrate buffer at pH 3.
So if we now just run our antibody through the block and then wash the block and elute off at pH 3, what you’re looking at here now is just a summary of the binding antibody as seen by analysis of the elution fraction, so this is the presence of our monoclonal in the elution fraction, so to be in the elution fraction it must have bound to the ligands. What you can see here is that some of these ligands do bind the antibody quite tightly and other ligands really don’t bind the antibody at all and we’ve got some that have partial binding. We scored these, just on a scale of 1 to 5 in terms of the ability to bind and we’ve colour-coded them and you’ll see a lot of this in the next examples just to make it easier to understand whether any given ligand is binding or not to the target. If we focus just in one area here, so this is zone 2 of the plate, so this includes ligands that are anionic in character and that have an aliphatic group, so they’re anions and aliphatic groups combined together in that zone.
If we take those fractions and run them out on an SDS-page gel we can see here that a number of these ligands are doing a very good job of binding impurities, host-cell protein, so we can look at lanes 5, 9 and 13, you can see that those ligands are pulling out a lot of protein that isn’t product related, this is host cell protein. If we look at other lanes, such as lane 6 and 10 and 14 you can see that those ligands are binding some host cell protein but also some antibody related impurities. So there’s a band here that at around 45 kilodaltons, that is a breakdown product of the heavy chain and we can see the excess light chain at around 25 kilodaltons and so some of these ligands are not binding the antibody, the whole antibody is much higher up the gel narrative under the 50 kilodaltons. Some of these ligands are not binding the whole antibody but they are binding host cell protein and they are binding the antibody related impurities. So they’re quite interesting ligands and we will be looking at those in a bit more detail in the next few slides.
We can also screen these libraries for host cell protein, so now in this case we’ve taken those elution fractions and we’ve screened them using a HCP assay, it’s a HTR FSA that we’ve used to estimate the concentration of host protein in there and you can see rows C and D in particular those ligands are lighting up quite strongly for the binding of host cell protein.
We can now compare that HCP binding data with our antibody binding data and if we just look at these now, so in the top left this is the slide you have seen previously for the elution fractions from the Multi-Mode Mimetic Ligand™ Library for the whole antibody, top right this is the HCP binding data that we’ve just seen, bottom left this is the binding of that 25 kilodalton antibody related impurity and bottom right this is the binding of the light chain, so this is the concentration of a light chain in the elution fraction. If we compare all of these different sets of data here, if we look now in yellow, these boxes represent ligands that have a high binding of HCP and a low binding of antibody and antibody related impurities. So if we wanted a polishing step that didn’t bind the antibody or any part of the antibody, but did bind the HCP then those ligands will be particularly good ligands to select and use for HCP removal from that particular antibody. Also, if we now look at the ones highlighted in orange, these are ones that also don’t bind whole antibody very well, they have maybe a small amount of binding for HCP but not much, but what they do have a very good binding of is the 45 kilodalton impurity and the light chain. So if we wanted a resin that was going to remove antibody related impurities, then those ligands highlighted in orange there would be a good choice for that application.
If we just look now in a bit more detail at those ligands that we’ve highlighted in yellow and in orange, we can see those here, so in lanes 3, 4 and 5 this is the SDS-page of those elution fractions and you can see there the presence of a large amount of host cell protein, of course HCP isn’t a single protein, it’s a wide spectrum of proteins that are related to the host cell organism, so it’s the cytosolic proteins typically from those cells. In lanes 6, 7 and 7, these were the orange ligands that are binding predominantly product-related impurities. In lane six there is actually ligand D9, that’s a particularly interesting ligand because as you can see, not only does it bind product-related impurities quite well, it also binds HCP quite well as well, so that’s almost a universal impurity binding ligand and in all of those ligands they have very low binding of the target protein, so the full antibody is passing through these columns and retarded. So this is an example of how we can use the Multi-Mode Mimetic Ligand™ Library to select a ligand, an adsorbent for the polishing of antibodies for the removal of either HCP or product-related impurities or both. Once a ligand has been selected from that screening, it’s been a week, and then contact Prometic, we can then provide larger amounts of any one of these ligands either a pre-packed column format in our EvolveD™ columns or we can supply any amount of slurry and we can provide large volumes of these in process applications as well, and we can provide the full support going forward. But the starting point, the identification of the adsorbent is what we are using this block for and it’s very quick and very efficient for doing that.
I’d like to move on now to some other examples, we are using now human plasma, this is partly because Prometic do a lot of work with human plasma and also because it’s a very complex feedstock, there are tens of thousands of different proteins in human plasma and it’s a very good source of proteins and you could really see the differences in selectivity that the ligands have and how we can use these ligands to pull out and remove different proteins from Plasma. So we would start by just looking at general binding of plasma proteins so in this example we are starting with equilibrating at pH 7.5 and we’re using a citrate buffer containing calcium chloride, and that’s because we don’t want the plasma to clot when it passes through the column, it’s whole plasma so we have all of the clotting factors here and fibrinogen, so if we mistreated that material it would form fibrin and it would block the column. We then load under those conditions, we’re loading a human source plasma to each column, we’re then washing with equilibration buffer and then we’re eluting here with this combination buffer that combines 1 molar sodium chloride with 10% hexanediol, and this is a really useful buffer for eluting protein from these multi-mode ligands that combine both charged an hydrophobic groups, that elution buffer contains salt to neutralise charge groups and it contains hexanediol which is very good, it’s quite soluble at 10% concentration but it’s good for reducing the polarity of the buffer to neutralise hydrophobic interactions.
So on the next slide we have now, the SDS-page of the elution fractions from the different columns in zones 2, 3, 4 and 5, you can see here just by looking that each one of these has a very different banding pattern, so if we look at, for example, lane 10 there you can see that ligand is binding albumin quite strongly, that large dark band. But if you look at rows 11 and 12 for example, those ligands that bind very little albumin. So each ligand has it’s own protein binding selectivity, each ligand will bind a different spectrum of proteins it’s just a question of screening them to determine which ligand binds the protein or proteins that we’re interested in.
On the next slide here we can see again these heat maps of binding where we’ve been looking at specific plasma proteins so at the top there these are ligands that we’ve identified that bind fibrinogen very well and the next one, the middle one, these are ligands that bind IVIG so this is polyclonal IgG, you see a different spectrum of ligands lighting up here to the ones that we found for monoclonal antibodies, so there is a difference here obviously between human plasma IgG and the monoclonals. Then at the bottom there, these are ligands for albumin binding, and if you recall a few minutes ago, I presented some data on the computational chemistry modelling and there was a prediction there that ligand C11 would be a good binder of albumin and as you can see here now, this is an actual screen for albumin and ligand C11 is one of the ones that is lighting up very strongly as a good albumin binder, so the prediction matches exactly what we see in practice with that ligand C11.
We can also use these ligands as well as binding individual proteins, we can look at the removal of unwanted impurities, now in the case of plasma proteins, impurities are actually human plasma proteins, so some of these are useful proteins that we want to have, but there can be cases where there are proteins in human plasma that we actually want to remove in particular where we’ve had activated proteases or activated clotting factors because these can, in the context of plasma protein therapeutics, these can have activated proteases or activated clotting factors that can give rise to either excessive bleeding or unwanted clotting and both of those are very undesirable for a therapeutic product.
So, there are situation where we want to specifically remove certain types of plasma protein impurities, so in the next few slides you’ll see how we’ve Multi-Mode Mimetic Ligand™ Libraries to do just that. So, we’ll start with an example of the capture and removal of IgA, so this is Immunoglobulin A, there are pharmacopoeia limits for the amount of IgA in IVIG products, excessive amounts of IgA are undesirable and it’s important that the IgA is removed. So, what we’ve done here is we’ve taken IVIG, we’ve spiked IgA into it and then we’ve used an IgA ELISA to look at the removal of IgA. So this is now looking at flow-through fractions and you can see that some of these are capturing IgA and some of them are letting IgA through in the flow-through. Some of the ligands here have very low binding of IgA in the flow-through, that means they are capturing IgA exceedingly well, so if you wanted a polishing step for IgA these would be the ones that we’d be interested in and that we’ve highlighted six ligands here that are able to capture IgA and all six of these have very low binding of IVIG, so more than 95% of the IVIG goes through in the flow-through fraction and the majority of the IgA is captured and bound and removed.
In another example, this is now the removal of Kallikrein, so Kallikrein is a protease when it’s activated, and it can catalyse the conversion of Plasminogen to Plasmin which is a very potent protease and that can lead to bleeding episodes, uncontrolled bleeding if there’s activated Kallikrein in Plasma product. So again, we’ve done a similar thing here, we’ve taken whole Plasma, we’ve applied a Kallikrein ELISA assay to the flow-through fractions, and the ones in red are the ones that are removing Kallikrein, a very low concentration of Kallikrein in the flow-through fraction because it’s being bound by the column and in this example we can see one ligand here in position G1 that was quite effective in capturing and removing Kallikrein but it doesn’t bind other, it has low binding of Plasma proteins particularly IVIG.
Another example here is Factor XI, this is a very real concern amongst plasma fractionators, there have been cases a few years ago of very large-scale product recalls of IVIG because of excessive activated Factor XI in the IVIG. So what we’ve done here is we’ve taken plasma and we have now a Factor XI ELISA and like we’ve done previously we’ve passed those through the column blocks and we’ve looked at the flow-through fractions and identified fractions that are low in Factor XI because that means we then capture the Factor XI on those columns. You can see here 3 ligands are very effective at the capturing and removing of Factor XI but they have low binding of IVIG which is what we want to see. So we took one of those ligands, that’s the ligand in the G1 position and we did a slightly larger scale experiment, so this is in a conventional packed column now, 2.5ml column, so we loaded our plasma onto that column, washed it through, you can see very large flow-through creeks as you kind of expect to see and then we eluted off the bound protein and you can see that quite sharp elution peak there at the point. The Factor XI is all contained in that elution peak, there’s very little Factor XI in the flow-through fraction.
If we now look at our SDS-Page of these fractions in lane 5 and 9, this is showing the elution fractions of these columns, so they have bound, not just Factor XI, they have bound a few other proteins, but if we look at the major plasma proteins like Albumin for example, IVIG, most of those are coming through in the flow-through fractions. So this particular ligand here, this ligand G1 is very effective in removing Factor XI from plasma or from IVIG that may have residual Factor XI present.
So to summarise now, what we’ve seen here is that we can use both affinity chromatography as a capture step and then use these Multi-Mode Mimetic Ligands™ as polishing steps and we can do that very effectively if we have a library of diverse multi-mode ligands. So Prometic has developed, just that with the Multi-Mode Mimetic Ligand™ Library, so this contains a very wide range of ligands that combine both ionic and hydrophobic groups and this can be rapidly screened to identify ligands for either capturing or polishing steps. So this is really a change of paradigm now for multi-mode ligands, so rather than spending a lot of time and effort in the lab to optimise the feed material so that we have binding of either the target or the impurities, what we now do is simply take the feedstock as it is, run it through the columns and select the ligand that provides the binding properties that we desire.
Finally, I’d just like to acknowledge the fact that a lot of work has gone into the discovery and selection of these ligands and the developments of the column block, some of my co-workers are listed, so Chloe, Graham, Buzz and Lucy, a big thank you to them and to the PBL research team.
At this point I will hand you all back to Lauren for any questions you might have, thank you.
Lauren Berry (BPI host): Thank you Steve 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 the 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 Prometic Bioseparations for sponsoring the event. Next, I’d like to quickly draw your attention to our face-to-face BPI conferences, visiting Tokyo, Santa Clara, Vienna and Boston next year, at these leading annual bioprocessing events you can hear from hundreds of expert speakers and network with peers. Also be sure to check out the resource list to the right of your screen where you can download the presentation slides from this session and a large collection of useful resources. Now back to Steve to begin the Q&A.
Question & Answer:
Lauren Berry (BPI host): So Steve we have a question here, do you have a ligand or ligands which specifically remove or reduce CHO or HCP in a flow-through mode for Mab of XC fusion protein products?
Steve Burton: Yes, I mean one of the examples I showed there was exactly that, so if we look at, I am just going to select one of the slides here. So if we see this particular slide I showed earlier, if you look at row 6 there, that particular ligand is doing exactly that, so it’s removing HCP and it’s removing product-related impurities from a CHO antibody, so I think this is an IgG1 antibody producing CHO cell lines, so that’s an example of doing exactly that.
Lauren Berry (BPI host): Great, another question that we have here. If I get a hit on the plate, what’s the next step?
Steve Burton: The next step would be to contact us and let us know which ligand or ligands you are interested in and then we can provide either ready packed columns, larger columns to do work with. So we have a range of columns, 5ml volumes or 50ml volumes that are pre-packed, or we can supply slurry, so we can supply 25ml upwards of samples of that particular adsorbent so that you can do your own continuing and developmental work. We can supply large amounts right up to 1000L batches if that’s required.
Lauren Berry (BPI host): Great, alright, another question. Can I get the structure of any of the ligands?
Steve Burton: Yes, we do, I mean we don’t publicise these at the beginning, but if any users of the plate have identified a ligand that they’re particularly interested in and they want to know more about it, we’d be very happy to share the structure of that ligand with you. That will probably require some kind of confidentiality agreement depending on the type of information that you would be interested in, but certainly we can provide information on the ligand structures and if you’re interested in any specific ligand and you wish to take that forward in your process development work.
Lauren Berry (BPI host): That’s great. Do you supply suitable generic binding elution conditions?
Steve Burton: We do have a list of buffers that we recommend for each of the zones on the plate, I highlighted this sort of general buffer that we have that combines salt and hexanediol, that’s often a very good starting point but we have other methods that we recommend and the information would be supplied with the Multi-Mode Mimetic Ligand™ Library as to what buffers you can try for elution for ligands in each of the 6 zones.
Lauren Berry (BPI host): What if I find a ligand that binds by target but not at the capacity I need?
Steve Burton: That’s a good question, so generally speaking for impurity removal for polishing steps the capacity required isn’t very high because the amount generally speaking isn’t a large amount of those impurities present, certainly not as to an affinity capture step. But in some cases it can be an issue so if you see a ligand that is giving a very good purification that the capacity maybe could be improved then we can develop another version of that so we can take that particular adsorbent and we can do some investigations, for example we can increase the ligand loading on the resin to give an increase in capacity or we might look at a different support matrix, for example we could look at a smaller beard to give a higher surface area, or we could look at a base bead that might have larger pores if porosity was limiting the binding. So there are various things that we can do here to increase the binding capacity of these resins after it being selected.
Lauren Berry (BPI host): Alright, one last question we have here. How quickly could I get a full-scale batch of any adsorbent on the plate?
Steve Burton: Okay, well in terms of larger amounts for continuing investigations we can provide those very quickly, they are off-the-shelf. Depending upon the scale of the batch that was required, we can certainly supply litre quantities very quickly and we can provide hundred litre quantities within probably 10 to 12 weeks of that order being identified to us. So, we can scale these things up relatively quickly. If there was any optimisation required then that would take a little time in addition to those times scales, if it was just simply a larger amount of the same material that was in the plate then we could produce larger amounts of that relatively quickly. So if there was a need for supporting information, some people might wish to have for example 3 batches produced as a part of a validation program we can do that, again there will be a longer time-scale associated with that, but we can certainly provide some larger batches and validated batches going forward.
Lauren Berry (BPI host): Great, that’s all the time that we have for questions today, thank you Dr. Burton for a great session. If anyone submitted a question that wasn’t addressed, please keep in mind that the speaker will reach out to you directly. This session was recorded, you’ll receive a notification in 24 hours when the on-demand session is available for viewing, also don’t forget about all the downloadable resources in the resources box to the right of the slides. Before you log off, please take a moment to complete the feedback form so we can continue to improve your digital week experience. On behalf of KNect365 Life Sciences, have a great day.
1x Multi-Mode Mimetic Ligand™ Library
5x Multi-Mode Mimetic Ligand™ Library
The Multi-Mode Mimetic Ligand™ Library is available for purchase immediately via Prometic Bioseparations online store, please visit the following web page to purchase: https://www.prometicbioseparations.com/product/multi-mode-mimetic-ligand-library/