Streamlining Glycan Analysis: From Sample Preparation to Data Analysis
Greetings everyone and welcome to another genetic engineering and biotechnology news webinar series. Our presentation today is entitled streamlining glycan analysis from sample preparation to data analysis. I'm Jeff Bogle esketh technical editor for gen and I'll be the moderator for today's webinar presentation. Glycosylation has had a tremendous impact on the biopharmaceutical industry as a fundamental attribute to monitor during the development and production of biotherapeutics proteins and regulatory agencies now requiring in-depth characterization using state-of-the-art analytical methodologies investigators are constantly on the hunt for glycan preparation and analysis tools that will save them precious time and money. Let's meet our speaker for this gen. Webinar who will describe a stream. Land glycan analysis platform that begins with sample preparation and is followed by data collection and processing which includes the corporation of a glidecam library search tool for accurate structural identification bill. Ali is a senior scientist for the Water Corporation. Dr. Ali will describe methodology for sensitive glycan measurements to take advantage of a simplified high-throughput sample prep procedure utilizing the glyco works rapid for MS and glidecam kit before we get rolling in the bills presentation. I want to encourage everyone to submit questions for our Q&A session at the end of the presentation. We'll try to answer as many questions as we can so simply type your question into the Q&A box on the left-hand side of your screen and hit submit all right with all the particulars out of the way. Let's get our webinar started bill. The audience is listening. Hello everyone. I'd like to thank you for attending today's webinar entitled streamlining glycan analysis from sample preparation to data analysis. I'd like to start off. Talk with an outline will be going over well first discuss quite Oscillation basics why black oscillation is important and why we want to study it in the biopharmaceutical industry i'll then discuss rapping for ms a molecule that we developed at waters that's been available for about a year and a half now and rafi for ms really helps out with both ethyl our and ms detection.
I'll then discuss how we integrate ready for ms into an automated analytical workflow that includes a library search. The library search really helps out with being able to assign structures and compositions to our flr. Peaks I'll then conclude by presenting a real-world example or comparing innovator and biosimilar infliximab using this analytical workflow. I'd like to start off. By just showing the basic building blocks the basic monosaccharides that are used to construct n linked glycans and in. Today's webinar will be discussing. Only n-linked glycans. I'd like to point out a few of the important monosaccharides one of those being. Sukkos down in the lower left-hand corner and the N glycol in termina Cassatt over on the lower right-hand corner and glycol ilnur mimic acid is a non human monosaccharides so its presence on biopharmaceutical products can cause some difficulties. If and present in the human body we also have galactose which when linked via an alpha linkage with another galactose is also non-human and can cause some problems. If present on a biopharmaceutical product as well there are three main classes of end of end linked glycans. There's the high- class. The complex class and the hybrid class and link lichens are composed of a common title bios core which is kind of is composed of two click next and three mannose residues. The high mannose class of glycans extends the Chi Tobias core with only mannose monosaccharides. The complex class of glycans will extend the core with a glut neck which can then be further extended with galactose and further extended with a silac. Acid complex lichens can also be few cos elated on the core or on an outer arm complex. Lichens can also have what we call bisecting. Glick neck which is attached to the core of the complex Lycan hybrid glycans are a combination of complex and the high.
Manos types of glycans clock. Oscillation is an important modification of proteins if we think about the human body at least 50% of proteins are glycosylated and glycosylation can influence a number of properties of the particular protein. So it's not surprising that a number of biotherapeutics are also glycosylated. Glycosyl lichens can be attached and thematically to a protein through asparagine residues and we call these n-linked glycans or they going to be attached to serine or threonine residues and these are called o-linked glycans in this webinar. We're only going to focus on in link lichens in black. Oscillation of proteins is routinely characterized and monitored during the development and manufacturing of biopharmaceuticals. And as we've just discussed glycosylated influences a number of physical chemical properties bio therapeutic proteins. If we consider a map that we see on this slide and the levels of coffee consolation can modulate its ADCC and its levels of high- glycans can alter its serum half-life file a file elation is interesting because it can lead to an tire pro-inflammatory responses and on human groups like an glycol ilnur mimic acid or to galactose is connected via an alpha linkage can induce immunogenic responses so with all of these different properties that can be influenced by glycosylation. The glycan profile of a biopharmaceutical is important to monitor. And it's also a critical quality attribute and glycosylation must be defined as rigorously as possible this slide further highlights the importance of glycosylation for bio pharmaceutical products. Again just giving some more specific examples where sigh elevation can decrease the serum half-life of EPO from 2 hours down to 10 minutes like oscillation analysis can be a very challenging task for multiple reasons. Unlike proteins like oscillation is not template driven so for a single composition that is the collection of monosaccharides that make up a given glycan.
It's possible to have several different structures or how the monosaccharides are connected. Together for example mini glycans contain a few coats as we saw a couple slides ago. But fucose can be located on the core or an outer arm silac acids can also be connected with connected by an alpha 2 3 or an alpha 2 6 linkage most commonly and unlike proteins and peptides glycans aren't linear chains. They're very highly branched structures and adding to the complexity. Glycosylation is also very sensitive to environmental or biological changes so slight changes in our cell culture conditions or cell media can result in significantly different glycan profiles analytically glycans can be quite challenging canalize they're not really ideal analytes for lc-ms they don't have an appropriate chromophore for detection by optical methods like UV or fluorescence. They don't have a basic site for ionization. In the positive mode of mass spectrometry therefore glycans require extensive sample preparation and Derivatives ation so they can be successfully analyzed by lc-ms however many sample. Prep protocols are very time-consuming. They can range from anywhere before to five hours out to several days and they can be very tedious and require significant skill to be able to get reproducible results and on top of that we haven't even talked about analyzing glycan lc-ms data while many bioinformatic tools are widely available to automate protein analysis. And peptide mapping experiments. These tools aren't as widely available for glycan analysis as we'd like them to be so. Many scientists still have to examine their data and make structural assignments by hand and there can still be some ambiguities in this first pass a lot of times. They have to go back and conduct tandem. MS experiments and interpreting this data can be a bit tricky as well so to address these analytical challenges and fill in some of these gaps waters has made significant investments in the area of glycol analysis.
Well sample preparation and in developing bioinformatic tools as we discussed on the previous like lichens are very challenging molecules than to analyze again analytically. They're not the ideal analytes for optical or mass spec methods some glycans contain sciatic acids and they can be detected in the negative mode of ionization. But it's more difficult to studied neutral glycans in this way so this isn't really a universal method so to be able to detect all the glycans they do need to be derivatized and in general the derivatives ation methods that we have developed we can consider them being grouped into different types of buckets so we have one bucket that can enhance optical detection one bucket that can enhance mass spec performance and I've included a third bucket for quantitation by mass spec since they have their own unique molecules if we look at the optical bucket we see molecules like two amino benzoate or two a B to enter nelec acid or two AAA. That can help out with both UV and fluorescence detection another molecule that's widely used particularly in capillary electrophoresis with laser induced fluorescence. Detection is eight amino pyrene one three six tri so phonic acid or just apts for short. If we look at some of the things that can enhance mass spec performance we see poor methylation which has been widely used or hydrozite tags that can increase the glycans hydrophobicity or can impart some more basic groups to enhance the mass spec ionization efficiency but overall we can see there's not much overlap between the two buckets for optical and mass spec performance with the exception of brokenwood. It does help out with both optical and mass spec performance so this lack of a single molecule that can enhance both mass spec and optical detection and be able to write label glycans rapidly was really an area that we wanted to invest the time and effort into to be able to develop a molecule to be able to do all of that in innocent.
Step the molecule that we develop that can help out with both lore essence and mass spec detection along with labeling glycans in a rapid manner we call rampy for ms an RF. Ms takes advantage of our long history and rapid labeling of amino acids using our a cue floor tag. Thank You Flores shown in the upper right hand corner and we can see that it has a rapid labeling group that's the NHS carbon made and it has a Quinlan yield group that allows it to be very active in the fluorescence detection. If we take a look at rappy floor at the bottom of the slide we can see that right before is really a second generation of this molecule. Where we've gone in and we've added on a church roof and that has a nice gas phase basicity and it can pick up protons and really help out with the ionization efficiency of glycans one important thing to remember when using our FMS is that we are labeling glycosyl amines. We don't allow the glycans once they've been released to be converted back to a reducing to a reducing sugar with the hydroxyl groups. This type of chemistry doesn't work doesn't work well for those types of molecules but by working with glycol amines. We're able to take advantage of the rapid reaction kinetics of the NHS carb and made group the labeling of the released end link glycans is done in and on in an almost instant manner in just a few seconds although we recommend letting the reaction go for about five minutes following the reaction. We'll see a shift of approximately three hundred eleven point one six. Dalton's rapid flora mess labels glycosyl means almost instantly while the time investment. The labeling step has been minimized the upstream sample. Prep workflow is still very time consuming and it can still be very labor intensive so it's not uncommon to spend several hours up to several days preparing the glycans just to get to the labeling step and then the labeling again it can take several hours plus a cleanup so it's very important for us to ensure that the sample prep workflow prior to the labeling could be done on a much faster time scale to take advantage of the rapid labeling of our FMS so we set out to redesign the workflow and we've invested quite a bit of time in being able to do this and I'd like to discuss that with you now so it begins with 50 micrograms of a glycoprotein.
They can be a monoclonal. Antibody some other bio therapeutic but it also works just as well with any other glycoprotein we first we denature it at 90 degrees. C with 1% Rapa. Just that's a little bit higher than what we use for. Peptide mapping experiments and we do that for 5 minutes then we let the sample cool down to room temperature and then we add rapid pngs. F that we co market with New England Biolabs to our release and Link likens we can get complete de cos elation 5 minutes at 50 to 55 degrees Celsius. The fastq like oscillation is very important to help preserve the glycosyl mean and not allow it to be converted back to reducing end. Then we label the release glycans with our FMS and we recommend a five-minute reaction and then we do need to purify the glycans to remove excess labeling reagent and we use a micro hillock illusion plate and this takes about 10 minutes so depending on the number of samples the entire sample prep workflow from beginning to end can be accomplished in as little as 30 minutes. Although if you're going to be preparing 48 or say 96 samples the time investment in sample. Prep will be a bit longer. But it's not going to be too bad following the sample cleanup. We just dilute the samples down with some acid or nitrile and dimethyl formamide. We don't dry them down in this feedback which also saves time and then once we've diluted the samples down we can just go straight into the you PLC and inject and do the mass spec analysis with this new workflow we do have a robust method. That gives us high yield.
We're able to achieve complete deep glycosylation in only 5 minutes. We verified this through a number of different tests including in tech mass analysis and gel shift assays for monoclonal antibodies. We've looked at sub unit by lc/ms we also have a very efficient labeling procedure we can get yields of at least 95% and we do take a bit of a hit on the solid phase extraction but we're getting a total recovery in the area of 74 to 75 percent but for a solid phase extraction. I don't think this is really too bad. I think it's quite good. I've done FB in the past and had a lot more sample loss so I'm quite pleased with the 74 percent. So overall we're getting a total yield of about 73 to 75 percent so that's actually really good. The derivatives glycans are quite stable as this slide shows on the Left. We're comparing the flr traces of a sample that was subjected to you. PLC MS analysis immediately following the sample prep and on the bottom on the left hand side. We're looking at that same sample. That was analyzed three days later and these traces are virtually identical on the right hand. Side is a chart. That's showing the flr peak areas for several monoclonal antibody glycans and we can see if there are no changes in the peak areas after three days so we can guarantee that the derivative derivative glycans are stable for three days. But we can't extend this out a little bit further so maybe up to a week or so before we start to see any noticeable loss it's really important to have stable derivatives because if we prepare or say 48 or 96 samples. They can take a little bit of time to get everything analyzed by the UV lc-ms analysis but having stable derivatives. For a week. You know we were very confident that we won't see any compromised sample integrity over that time period to benchmark our FMS. We've compared to several other commonly used tags including instant a B to a B and pro Kahneman and here we're showing the response factors of the other labels as percentages of the response factor of our FMS and the gains in fluorescence and mass spec sensitivity are apparent in these two plots since it does portray the response factors for the other molecules normalized to that of our FMS on the Left we're comparing the performance of the flr and our FMS gives the highest signals within today be giving the next highest signal but it's only about half of what we get for our FMS on the right we're comparing the mass spec performance and again our FMS gives us the highest signals with pro Kadima this time coming in second place but it's only about a third of what we see for our FMS.
This is actually a very interesting result. Since both of these labels contain a tertiary amino group and we believe that the enhanced signal intensity of our FMS label glycans originates from our FMS having a bit more hydrophobic character than the pork and mid label and there are some studies out there that show that additional hydrophobic surface area for glycan does lead to increase electrospray ionization efficiency so our FMS is the first molecule on the market that can label glycans in an almost instant matter improve the fluorescence signal detection and improve the mass spec performance now let's take a look at how we can integrate our FS mislabelling into a streamlined analytical workflow that includes searching an RF IMS glycan geo scientific library to make the task of glycan analysis easier we've developed an application solution that streamlined data collection data processing and data analysis. This solution has three main components the first being hardware the second is chemistry and the third is software the hardware is our analytical system and includes an acuity you plch class bio system and that has an acuity you PLC column manager and acuity you PLC flr detector and a 0g to xsq top mass spectrometer. The chemistry includes the glycol works rapid floor. MS and glycan kit and an acuity you PLC glycan BHM and column the software that we use in this solution is our unified scientific information system and when we integrate all of the components into a single system.
We're able to streamline the process of data collection all the way through the reporting and this solution makes the task of glycan analysis. Much easier and faster. Now let's take a look at how unify works and in particular the glycan application unless there's a generic workflow within unify that acquisition processing the bioinformatics and reporting tools are all integrated and automated within the unified scientific information. System and unify is a compliant. Ready architecture in the middle is our hillock flr with accurate mass workflow. This solution allows the routine assignment of n-linked glycans structures using hillock you. PLC data from the flr that has been time aligned with the mass spec channel. The chromatographic data are then processed and calibrated and converted into glucose units. And we'll talk about glucose units for just in just a bit. The mass spec data are similarly processed with a lock mass calibration our hillock flr with accurate mass that are then searched against one of our geo scientific libraries for structural determination and the search feature simplifies and expedites the task of structural assignments unify also offers an alternative workflow for. Dda types of analyses. The data are acquired then processed and during the processing the data are d convoluted and D isotopes the file can then be converted and exported as LCS or as an MZ XML file and these files are compatible with the sim glycan software that can help automate the tasks of tandem ms structural characterization and I would just like to point out that the unified glycan workflow does support both to a B and rappy floor label glycans. It's important to note that the workflow that we recommend is a dual detector method and we do utilize both the flr and mass spec channels we use DSLR for glycan library searching and for quantitation and our quantitation.
Typically we use a normalization scheme where the flr peak area of a particular glycan is expressed as a percentage of the total peak area for all of the glycans identified by the library. Search we don't use the MS channel for quantitation we use that information to confirm our tu scientific library searches and I'd like to point out that our FMS has greatly helped out with our GU. With accurate mass workflow particularly for the lower buttons glycans and in this slide. I'm comparing the flr. Trace on the top to the mass spec base become a deterrent base peak intensity trace on the bottom and this shows really how closely the two detectors mirror each other and a particular interest are those glycans that later in the run and our president very low abundances as we can see in the inset using other tags we've really struggled to get high quality mass spec data to confirm the library search results but with our FMS. We're able to get enough of mass spec signal to be able to really confidently identify what those particular glycans are and as a general rule for me if I see a peak and the flr with an intensity of around 0.2. I'm quite confident that I'll get a good enough mass spec data to get that accurate mass confirmation of course this is a very rough estimate as all F all our detectors and bulbs are unique as our mass spectrometers. But again for me. This general rule works out quite well and so now. Let's talk about the our finesse. GU scientific library in a little bit more detail. This library is very similar to our to a bee library. So if you're comfortable using the - a bee library using the RF MMS library isn't going to be any problem the RF MF library was generated using nine different glycoproteins from several different biological sources. We use several monoclonal antibodies like etanercept infliximab trastuzumab and EPO. The nist monoclonal antibody pulled human IgG mouse IgG bovine. Fecha one for a source of highly branched in Sayulita glycans and yeast invert aids as a source for high mono species.
This library contains GU values for 177 glycans that we tailored to the biopharmaceutical industry and we realize this isn't an exhaustive library and some users may have some unique lichens that are in the library but unify allows users to create their own custom libraries that they can co search with the RFM s glycan GU scientific library. The library does contain all the major types of end link lichens it has high mannose glycans complex lichens and hybrid glycans. There's a wide variety of Sai allele ADA glycans including those that have n glycol ilnur Minik acid and and a settlor Minnick acid. There's also several glycans that have alpha linkle. Actos pairs and the library also contains GU values for many isomeric lichens including positional isomers and linkage isomers. So for example for several plugins we had we can distinguish alpha 2 3 links. Ayala Casas from alpha to 3 F from alpha to six links sciatic acids the development of the RF misc like in G scientific library with a joint project between Waters and Pauline Rudd and mark Hillard of the National Institute for bio processing research and training or nyebern Ireland and they're the ones that really didn't an exhaustive characterization of the glycans using a battery of EXO Casa de this for each bike and sample and using the EXO Casa de zzz's they were able to selectively remove specific monosaccharides and this gives us a very precise structural characterization. We gave a lot of information out of this. And this characterization is shown in the upper left-hand upper left hand side of this slide so for example they start off by running den digested sample and then they would first remove alpha to three linked silac acids and they could look at the shift in retention time in GU values and then they would remove the alpha to six linked silac acids and then they would remove you know subsequent monosaccharides so really doing a very thorough job of really pinning down what the exact structure of each glycan is and then each glycan has also been subjected to mass confirmation.
And for a lot of glycans. I've also collected mmmmm SMS data. And and. So yeah i'm very confident that all of the GU values and for the structures you know they are in fact correct. And why did we go through all of this effort of EXO class today's digestion z' and getting accurate mass data. It was really so that the users didn't have to do this. That all this work has already been done for you so you can just you know. Run your samples do the library search and get your results and I'd like to point out that each sample was analyzed at least ten times with the mass confirmation so we do have great confidence in the accuracy of the GU values since we are searching a library based on calibrated retention times. We do have a set of conditions for the separation that we do need to follow fairly closely to get the best library search results we do need to use the 2.1 by 150 millimeter acuity u PLC glycan b hm and column that has the 1.7 micron particles with a pore size of 130 angstroms. This is the same column. That's been used for the 2a b library so that part is the same the mobile phases mobile phase a is 50. Mm hola ammonium formate at ph 4.5 and mobile phase B is pure acid or nitrile with no modifier we conduct the separations for the RF MS experiments at 60 degrees C. It's bit higher than what we use for the 2a b library and we use a gradient from 75% b to 54% B over 35 minutes or a slope of zero point six percent per minute this gradient is a little bit different than the one we used for the two AP library but the column is the same. The mobile phases are the same just a couple of slight changes in the conditions from going from the to a/b experiments to the rapid flora math experiments. We use glucose units instead of raw retention times we found a GU values to be a little bit more stable across different instrument systems or across different columns or even across different days while retention times do tend to shift around a little bit and so if we search a library just based on retention times our search results may not be as well as we would like them to be so we recommend using GU values to be able to convert retention times into glucose units.
We first run an analysis using dextran has been labeled with rappy for rms and we can see this in the figure on the upper right hand part of the slide we then tell the software the exact retention time for the structures for example. We're telling the software that gu4 has a retention time of seven point. Five three minutes gu5 is ten point six six minutes and on and on and on we can then convert those GU values and retention times into a calibration curve shown just below the flr chromatogram and then we can generate an equation based on this data that we can then use to convert through tension times of our unknown samples into GU values which we can then use to search the scientific library in order to obtain the best GU values and get the most accurate scientific library search results for our FMS we do recommend using our rappy flora mess texture and calibration ladder it took a little bit of chemistry to be able to get our FMS attached to it but we were able to do that through an ethanolamine. Oh you're a linkage. And our dexter and ladder covers gu values from two out to thirty alignments to get accurate. G values from very small glycans out to very large ones. So you know some of these huge glycans that are you know highly branched with you know tetran ten ireri four silac acids of yukos's maybe throw in some like. Tosa mean extensions we can get accurate GU values for all of those types of glycans and to get the best library search results. We recommend running the calibration ladder with each analysis set and now let's take a look at some fully processed results that we get from unify and this slide is a recreation of what we would see on the left hand side.
We see the flr trace. That's been annotated automatically with the glycans from the library search result and if we click on any one of those structures we can pull out the extracted 2-iron chromatogram and here. We see two peaks when we do that so right away. I know that we're dealing with isomers. And if we want we can pull out the process mass spectra as well on the right hand side of the slide. We see the library search results and I'd like to explain how the library search works. We do search within a GU window we recommend plus or minus 0.2 GU units and after unify has calculated the GU value for a given peak in the flr chromatogram. It'll search the library for all of the glycans that fall within that window so for the highlighted peak in the flr trace unified determinate had a GU value of eight point one eight so the library search pulled out all of the glycans that had GU values between seven point. Nine eight and eight point three eight then. The software goes back and we'll determine which like and out of that pot out of that list is actually the correct one and this is where the mass spec data comes into play. It'll go back into the mass spec data it'll look for an ion and it can search different charge states. That's not an issue but it looked for an ion corresponding to that particular glycan in the mass spectra if it sees that ion with an acceptable acceptable mass accuracy about 2. Ppm then we say is that glycan is mass confirmed and if we look at the results for this particular search the glycan that's highlighted had a delta GU of 0.01 one GU units and a Delta M over Z of 0.005. So these are both very accurate results and so there's absolutely no question that the library has given us the correct glycan for that particular. GU value and now. Let's take a look at how we can use this workflow in a real-world study where we were comparing innovator and biosimilar inputs and map and can looking at the glycan profiles in flip.
The map has been on the market for about twenty years. Now it's an anti TNF alpha monoclonal antibody. It was originally approved to treat rheumatoid arthritis and it's now been approved to treat. Crohn's disease and a number of other diseases as well. There's been a biosimilar version known as in fly code has been developed by a company based out of Korea the biosimilar version has been approved for use in Canada Australia and South Korea and importantly this is the first biosimilar monoclonal antibody. That's been approved for use in the United States. And so we thought this would be a very timely comparison to do between the biosimilar and the innovator infliximab to begin our comparison of innovator and biosimilar infliximab we first looked at the flr profiles. We were able to obtain three samples of the innovator molecule in one sample of the biosimilar but if we look at the flr traces we can see how similar these four traces are. There are some slight differences with some of the lower buttons glycans. It appears but for the most part the major structures appear to be very similar we were able to identify 23. Unique lichens found the innovative molecule and 21 on the biosimilar. All of the glycans are on the biosimilar. Were also on the innovator molecule so there are two unique glycans on the innovator monoclonal antibody once we've done the library searching we've determined what the carbohydrates on the sample are we can start to group those into various classes based on their structural features. And this is what this slide is showing and we're just comparing different classes of glycans between the innovator molecules which are shown in blue and the biosimilar which is shown in red so we're able to classify in this example sugars that are neutral mono sila violated nazi. Consulate's be conflated mono and generate binary high menos and those that have alpha and K lactose units and I think overall we can see that the different classes are very similar.
There appear to be some slight differences for example it looks like the neutral glycans are slightly more elevated in the innovator sample and perhaps a mono thiolated are slightly more abundant in the biosimilar sample but overall we think that the the different classes are present at a very very similar abundance levels on both on both types of molecule to drill down into the data and really look at specific glycans that are different in their abundance levels. Between the innovator in the biosimilar molecule we can start to look at these mirror image plots and in this prodding at the biosimilar shown on top in green and the innovator shown on the bottom in black and the difference in the are on density of these peaks is shown as the red trace in the middle. Now we are looking at raw data in this example but we believe that since we start with the same amount of starting material 15 micrograms of each monoclonal antibody and we inject the same amount. We've already taken care of there on those agents. We believe that this is a valid comparison of things. If we look at speak at the retention time of 12.94 minutes we need to rotate it. Dips down then jumps back up which is indicating that there are no differences in the peak intensity for that particular models for that particular carbohydrate if we zoom in on some of the carbohydrates that are lower their abundance levels absolute later around 18 minutes to about 24 minutes. We can start to see. There are some some differences those abundance levels the interesting thing is that these are all terminated within glycol in Germanic absolute and they all appear to be more elevated and their abundance bubbles on the biosimilar monoclonal antibody although it is worth noting that these are present at very low abundance levels when we start to look at all of the innovator monoclonal antibodies we can start to see those patterns those on the mirror image.
Plots starting to really be emphasized for the glycans that are capped with and glycol aromatic acid and overall looks like that those types of glycans are more abundant on the biosimilar molecule. We have four examples of those in general. It looks like these dilated species were three times more abundant on the biosimilar molecule as on the innovator molecule if we look at glycans that have alpha linked galactose pairs. Those tend to be more prevalent. On the innovator molecule in fact this hybrid glycan shown on the bottom right hand corner was only observed on the innovator molecule an interesting lichen was the one that had both an acetylene or minik acid and alpha-linolenic. What i find interesting about this was it was more abundant on the on the biosimilar molecule although it wasn't as abundant as some of those some of the other carbohydrates found that didn't have the alpha galactose linked to it so. I think it's really interesting when we look at these glycans that can have some immunogenic responses that there seems to be a preferential expression for the for the type of map again the the glycans that have n glycol aromatic acid tend to be more abundant on the biosimilar. The glycans that have alpha like galactose pairs appear to be more abundant. On the innovator molecule it's also important to take a look at the levels of yukos and again we can start to look at that those levels once we've done the library search over all the levels of fucose were quite similar on both molecules there was one particular glycan that we call. FA - six G for one that appeared to be more abundant on the innovator monoclonal antibody but that increase was kind of compensated for by the s6 a tube-like and so overall the levels of the constellation on both types of map were very similar we concluded I'd like to highlight some of the key points that we discussed. We're very excited about right before. M s this is the first molecule on the market that can rapidly or instantly label.
What console amines that can enhance both the flr and the mass spec signal again. This reaction is done in a matter of seconds but we recommend letting it go for a few minutes. We can then integrate this into a simple sample preparation workflow. We've been able to cut the time investment down from several hours to a few days down to just a few minutes so we can. We can get the sample preparation done in as little as 30 minutes. But depending on the number of samples it may take up to an hour we can then incorporate this simplified sample preparation scheme into a simplified data acquisition and analysis within unify. We're very excited about being able to simplify the data analysis using our scientific library search glycan values traditionally. This has been an area. That's been very difficult and can be very time-consuming but with the ability to search a scientific library greatly simplifies this task and it makes it more straightforward and a lot faster we can then take this overall workflow of FRA performance labeling with our automated data acquisition and library searching we can start to develop some very interesting applications and in the example that I showed today we can start to compare monoclonal. Antibodies from innovator and biosimilar companies and we're able to look at what types of glycans are different in what types of glycans are the same. And so with that there's a and with that there are several people that I'd like to thank I would like to thank Pauline Rudd and market leaders. Nyberg they've been instrumental in assisting with the core development of the RF. Ms like in geo scientific library. I'd also like to thank Matthew Lauber Yin Jing to ROS Waller Scott Berger MN do Jennifer Fournier and Alexei Armstrong. The waters for helping put this presentation together. I'd like to take this opportunity to thank Jeff for giving us this opportunity to present some of the research that we're doing up waters in the area of glycan analysis.
We're very excited about some of these developments that we believe will really streamline the task of glycan analysis. Thanks bill that was a great presentation you provided audience with some great information about the advantages of the rapid for MS kit and how it will streamline their glycan analysis workflows so we thank you for that before we get started on the Q&A session. I want to remind everyone once again that this is your final chance to submit questions for our speaker so hurry up and submit them now just type them. In to the box to the left-hand side of your screen and hit submit all right to have some really great questions that have already started to roll in. So why don't we get to answering as many of them as we can bear with us for a few moments while we get everything situated on our end and we'll begin the Q&A all right everyone thanks for joining us for the Q&A have a few questions that have rolled in already. So let's get to them. Bill first question is. Can we add glycans to the library. In addition to the more glycans. I have some very unique structures and I want to be sure that I can use the library as effectively as possible. Yeah that's a great question. Unfortunately we can't edit glycans that are in the library. We've got that locked down but what we can do within unify is we can create our own custom library so for the customers that are dealing with some unique structures that they've engineered for their particular protein. We can go in. We can create that like lichen structure. We can calculate the data and we can create a unique library. Put that glycan in that library and then we can go search with the with the rms geo glycan library and we'll be able to get the structure identified that way but but we can't edit the actual library itself and go searching. The library works just fine. Works great alright so our next question one of our audience members would like to know. Can we use the library outside of unifying.
I'm using our for glide analysis and would like to be able to use this library. I yes you can. This was co-developed with with knivert. And this has all been uploaded into their glyco base database. So you can get this exact same data through glycol base. They have to register for an account with that. But you can search the data it is manual. It's not automated like what we would do and unify but you do have access to that data outside of unify. Yes okay great. Another question audience member starts off by saying. I'm working with some very large glycans that elute very late in the lc/ms analysis. How much flick is there with the PSU PLC condition. We do have a little bit of flexibility particularly with what the length of the gradient so we run a gradient from 75 percent zero nitrile to 54 percent of cedar nitrile if you need to drop that down to say 45 percent or 40 percent that's fine that's not going to affect anything you can you can lengthen the gradient but we do need to keep the gradient at the 0.6% per minute what we can't do is change the temperature. If we start to change the temperature then we start to change the selectivity of the column and then our library searches aren't artists as as accurate as we'd like them to be but lengthening the gradient not a problem if you want to shorten it a little bit. That's not going to be a problem either. Just make sure that we keep the slope at 0.6% per minute and and but we'll get very accurate. Library searches. Thank You boom another question. Will the decode will the deck like oscillation step eventually give a monosaccharide or was it a mixture of mono and oligosaccharides the dig like. Oscillation step we're just using the rapid beam jeiza from New England Biolabs and it you know it's the access as normal P&G ASF does and it will release intact and link likens as glycosyl amines. So yeah we don't release those as monosaccharides. Alright thank you we have one person who will show. Kaiba and I just wanted to mention the one thing that we we have done is we have optimized the buffer with the rapid P and JSF so that it does work better with rappy floor ms so the buffer is a little bit different.
That's why we're cool marketing with with. New England Biolabs. Alrighty thank you so water. Navarro I guess missed a little bit information and would like to know if you could repeat. How do you maintain the glycosyl mean during PNG ACF treatment and what keeps it from converting to reducing it. Yep that's a great question so the ability to to do that the D glycosylation is we do use quite a bit of high concentration of the P and GS F and we do the reaction in five minutes. And the the half-life of the glycol amine is a couple of hours so by getting the digestion done in five minutes and then doing all the subsequent sample of derivatives ations. You know in another five minutes or so. You know we're able to get the we're able to traffic lichens as glycosylated efficiently. Before they have a chance to be converted back down to the to the hydroxyl so really we're just taking advantage of the rapid kinetics and being able to wrap it too rapidly label things so we're just taking advantage of the time okay and it looks like we have our last question. Audience member would like to know. How do you determine the stereochemistry of glycosidic linkages it based solely on the glycan library. Yes so this is all ties back into the very rigorous characterization with the expo galactosidase digestion that. There are co-workers and collaborators that. Nyberg did or they're using very specific XO glycolysis to be able to you know very selectively remove amino acids based on you know alpha beta linkages. And you know alpha to 3 vs alpha to six linked sciatic acids and you know the the galactose linkages as well and and on and on and on and on down for each each monosaccharides so we're getting that type of information from our very rigorous characterization using very specific XO qui casa dayses alrighty.
Well thank you bill and with that. It looks like we've come to the end of our webinar so want to remind everyone that our webinar will be archived for up to a year on our website at WWE and it you can watch it again or feel free to for the link to your friends and colleagues which we always recommend. I'd like to thank Bill again for his informative presentation and I like to thank the audience for their attention and thoughtful questions and the very special thanks to Waters for sponsoring this webinar. So hopefully we'll see you again at another gen webinar in the near future. Good bye for now.
I'll then discuss how we integrate ready for ms into an automated analytical workflow that includes a library search. The library search really helps out with being able to assign structures and compositions to our flr. Peaks I'll then conclude by presenting a real-world example or comparing innovator and biosimilar infliximab using this analytical workflow. I'd like to start off. By just showing the basic building blocks the basic monosaccharides that are used to construct n linked glycans and in. Today's webinar will be discussing. Only n-linked glycans. I'd like to point out a few of the important monosaccharides one of those being. Sukkos down in the lower left-hand corner and the N glycol in termina Cassatt over on the lower right-hand corner and glycol ilnur mimic acid is a non human monosaccharides so its presence on biopharmaceutical products can cause some difficulties. If and present in the human body we also have galactose which when linked via an alpha linkage with another galactose is also non-human and can cause some problems. If present on a biopharmaceutical product as well there are three main classes of end of end linked glycans. There's the high- class. The complex class and the hybrid class and link lichens are composed of a common title bios core which is kind of is composed of two click next and three mannose residues. The high mannose class of glycans extends the Chi Tobias core with only mannose monosaccharides. The complex class of glycans will extend the core with a glut neck which can then be further extended with galactose and further extended with a silac. Acid complex lichens can also be few cos elated on the core or on an outer arm complex. Lichens can also have what we call bisecting. Glick neck which is attached to the core of the complex Lycan hybrid glycans are a combination of complex and the high.
Manos types of glycans clock. Oscillation is an important modification of proteins if we think about the human body at least 50% of proteins are glycosylated and glycosylation can influence a number of properties of the particular protein. So it's not surprising that a number of biotherapeutics are also glycosylated. Glycosyl lichens can be attached and thematically to a protein through asparagine residues and we call these n-linked glycans or they going to be attached to serine or threonine residues and these are called o-linked glycans in this webinar. We're only going to focus on in link lichens in black. Oscillation of proteins is routinely characterized and monitored during the development and manufacturing of biopharmaceuticals. And as we've just discussed glycosylated influences a number of physical chemical properties bio therapeutic proteins. If we consider a map that we see on this slide and the levels of coffee consolation can modulate its ADCC and its levels of high- glycans can alter its serum half-life file a file elation is interesting because it can lead to an tire pro-inflammatory responses and on human groups like an glycol ilnur mimic acid or to galactose is connected via an alpha linkage can induce immunogenic responses so with all of these different properties that can be influenced by glycosylation. The glycan profile of a biopharmaceutical is important to monitor. And it's also a critical quality attribute and glycosylation must be defined as rigorously as possible this slide further highlights the importance of glycosylation for bio pharmaceutical products. Again just giving some more specific examples where sigh elevation can decrease the serum half-life of EPO from 2 hours down to 10 minutes like oscillation analysis can be a very challenging task for multiple reasons. Unlike proteins like oscillation is not template driven so for a single composition that is the collection of monosaccharides that make up a given glycan.
It's possible to have several different structures or how the monosaccharides are connected. Together for example mini glycans contain a few coats as we saw a couple slides ago. But fucose can be located on the core or an outer arm silac acids can also be connected with connected by an alpha 2 3 or an alpha 2 6 linkage most commonly and unlike proteins and peptides glycans aren't linear chains. They're very highly branched structures and adding to the complexity. Glycosylation is also very sensitive to environmental or biological changes so slight changes in our cell culture conditions or cell media can result in significantly different glycan profiles analytically glycans can be quite challenging canalize they're not really ideal analytes for lc-ms they don't have an appropriate chromophore for detection by optical methods like UV or fluorescence. They don't have a basic site for ionization. In the positive mode of mass spectrometry therefore glycans require extensive sample preparation and Derivatives ation so they can be successfully analyzed by lc-ms however many sample. Prep protocols are very time-consuming. They can range from anywhere before to five hours out to several days and they can be very tedious and require significant skill to be able to get reproducible results and on top of that we haven't even talked about analyzing glycan lc-ms data while many bioinformatic tools are widely available to automate protein analysis. And peptide mapping experiments. These tools aren't as widely available for glycan analysis as we'd like them to be so. Many scientists still have to examine their data and make structural assignments by hand and there can still be some ambiguities in this first pass a lot of times. They have to go back and conduct tandem. MS experiments and interpreting this data can be a bit tricky as well so to address these analytical challenges and fill in some of these gaps waters has made significant investments in the area of glycol analysis.
Well sample preparation and in developing bioinformatic tools as we discussed on the previous like lichens are very challenging molecules than to analyze again analytically. They're not the ideal analytes for optical or mass spec methods some glycans contain sciatic acids and they can be detected in the negative mode of ionization. But it's more difficult to studied neutral glycans in this way so this isn't really a universal method so to be able to detect all the glycans they do need to be derivatized and in general the derivatives ation methods that we have developed we can consider them being grouped into different types of buckets so we have one bucket that can enhance optical detection one bucket that can enhance mass spec performance and I've included a third bucket for quantitation by mass spec since they have their own unique molecules if we look at the optical bucket we see molecules like two amino benzoate or two a B to enter nelec acid or two AAA. That can help out with both UV and fluorescence detection another molecule that's widely used particularly in capillary electrophoresis with laser induced fluorescence. Detection is eight amino pyrene one three six tri so phonic acid or just apts for short. If we look at some of the things that can enhance mass spec performance we see poor methylation which has been widely used or hydrozite tags that can increase the glycans hydrophobicity or can impart some more basic groups to enhance the mass spec ionization efficiency but overall we can see there's not much overlap between the two buckets for optical and mass spec performance with the exception of brokenwood. It does help out with both optical and mass spec performance so this lack of a single molecule that can enhance both mass spec and optical detection and be able to write label glycans rapidly was really an area that we wanted to invest the time and effort into to be able to develop a molecule to be able to do all of that in innocent.
Step the molecule that we develop that can help out with both lore essence and mass spec detection along with labeling glycans in a rapid manner we call rampy for ms an RF. Ms takes advantage of our long history and rapid labeling of amino acids using our a cue floor tag. Thank You Flores shown in the upper right hand corner and we can see that it has a rapid labeling group that's the NHS carbon made and it has a Quinlan yield group that allows it to be very active in the fluorescence detection. If we take a look at rappy floor at the bottom of the slide we can see that right before is really a second generation of this molecule. Where we've gone in and we've added on a church roof and that has a nice gas phase basicity and it can pick up protons and really help out with the ionization efficiency of glycans one important thing to remember when using our FMS is that we are labeling glycosyl amines. We don't allow the glycans once they've been released to be converted back to a reducing to a reducing sugar with the hydroxyl groups. This type of chemistry doesn't work doesn't work well for those types of molecules but by working with glycol amines. We're able to take advantage of the rapid reaction kinetics of the NHS carb and made group the labeling of the released end link glycans is done in and on in an almost instant manner in just a few seconds although we recommend letting the reaction go for about five minutes following the reaction. We'll see a shift of approximately three hundred eleven point one six. Dalton's rapid flora mess labels glycosyl means almost instantly while the time investment. The labeling step has been minimized the upstream sample. Prep workflow is still very time consuming and it can still be very labor intensive so it's not uncommon to spend several hours up to several days preparing the glycans just to get to the labeling step and then the labeling again it can take several hours plus a cleanup so it's very important for us to ensure that the sample prep workflow prior to the labeling could be done on a much faster time scale to take advantage of the rapid labeling of our FMS so we set out to redesign the workflow and we've invested quite a bit of time in being able to do this and I'd like to discuss that with you now so it begins with 50 micrograms of a glycoprotein.
They can be a monoclonal. Antibody some other bio therapeutic but it also works just as well with any other glycoprotein we first we denature it at 90 degrees. C with 1% Rapa. Just that's a little bit higher than what we use for. Peptide mapping experiments and we do that for 5 minutes then we let the sample cool down to room temperature and then we add rapid pngs. F that we co market with New England Biolabs to our release and Link likens we can get complete de cos elation 5 minutes at 50 to 55 degrees Celsius. The fastq like oscillation is very important to help preserve the glycosyl mean and not allow it to be converted back to reducing end. Then we label the release glycans with our FMS and we recommend a five-minute reaction and then we do need to purify the glycans to remove excess labeling reagent and we use a micro hillock illusion plate and this takes about 10 minutes so depending on the number of samples the entire sample prep workflow from beginning to end can be accomplished in as little as 30 minutes. Although if you're going to be preparing 48 or say 96 samples the time investment in sample. Prep will be a bit longer. But it's not going to be too bad following the sample cleanup. We just dilute the samples down with some acid or nitrile and dimethyl formamide. We don't dry them down in this feedback which also saves time and then once we've diluted the samples down we can just go straight into the you PLC and inject and do the mass spec analysis with this new workflow we do have a robust method. That gives us high yield.
We're able to achieve complete deep glycosylation in only 5 minutes. We verified this through a number of different tests including in tech mass analysis and gel shift assays for monoclonal antibodies. We've looked at sub unit by lc/ms we also have a very efficient labeling procedure we can get yields of at least 95% and we do take a bit of a hit on the solid phase extraction but we're getting a total recovery in the area of 74 to 75 percent but for a solid phase extraction. I don't think this is really too bad. I think it's quite good. I've done FB in the past and had a lot more sample loss so I'm quite pleased with the 74 percent. So overall we're getting a total yield of about 73 to 75 percent so that's actually really good. The derivatives glycans are quite stable as this slide shows on the Left. We're comparing the flr traces of a sample that was subjected to you. PLC MS analysis immediately following the sample prep and on the bottom on the left hand side. We're looking at that same sample. That was analyzed three days later and these traces are virtually identical on the right hand. Side is a chart. That's showing the flr peak areas for several monoclonal antibody glycans and we can see if there are no changes in the peak areas after three days so we can guarantee that the derivative derivative glycans are stable for three days. But we can't extend this out a little bit further so maybe up to a week or so before we start to see any noticeable loss it's really important to have stable derivatives because if we prepare or say 48 or 96 samples. They can take a little bit of time to get everything analyzed by the UV lc-ms analysis but having stable derivatives. For a week. You know we were very confident that we won't see any compromised sample integrity over that time period to benchmark our FMS. We've compared to several other commonly used tags including instant a B to a B and pro Kahneman and here we're showing the response factors of the other labels as percentages of the response factor of our FMS and the gains in fluorescence and mass spec sensitivity are apparent in these two plots since it does portray the response factors for the other molecules normalized to that of our FMS on the Left we're comparing the performance of the flr and our FMS gives the highest signals within today be giving the next highest signal but it's only about half of what we get for our FMS on the right we're comparing the mass spec performance and again our FMS gives us the highest signals with pro Kadima this time coming in second place but it's only about a third of what we see for our FMS.
This is actually a very interesting result. Since both of these labels contain a tertiary amino group and we believe that the enhanced signal intensity of our FMS label glycans originates from our FMS having a bit more hydrophobic character than the pork and mid label and there are some studies out there that show that additional hydrophobic surface area for glycan does lead to increase electrospray ionization efficiency so our FMS is the first molecule on the market that can label glycans in an almost instant matter improve the fluorescence signal detection and improve the mass spec performance now let's take a look at how we can integrate our FS mislabelling into a streamlined analytical workflow that includes searching an RF IMS glycan geo scientific library to make the task of glycan analysis easier we've developed an application solution that streamlined data collection data processing and data analysis. This solution has three main components the first being hardware the second is chemistry and the third is software the hardware is our analytical system and includes an acuity you plch class bio system and that has an acuity you PLC column manager and acuity you PLC flr detector and a 0g to xsq top mass spectrometer. The chemistry includes the glycol works rapid floor. MS and glycan kit and an acuity you PLC glycan BHM and column the software that we use in this solution is our unified scientific information system and when we integrate all of the components into a single system.
We're able to streamline the process of data collection all the way through the reporting and this solution makes the task of glycan analysis. Much easier and faster. Now let's take a look at how unify works and in particular the glycan application unless there's a generic workflow within unify that acquisition processing the bioinformatics and reporting tools are all integrated and automated within the unified scientific information. System and unify is a compliant. Ready architecture in the middle is our hillock flr with accurate mass workflow. This solution allows the routine assignment of n-linked glycans structures using hillock you. PLC data from the flr that has been time aligned with the mass spec channel. The chromatographic data are then processed and calibrated and converted into glucose units. And we'll talk about glucose units for just in just a bit. The mass spec data are similarly processed with a lock mass calibration our hillock flr with accurate mass that are then searched against one of our geo scientific libraries for structural determination and the search feature simplifies and expedites the task of structural assignments unify also offers an alternative workflow for. Dda types of analyses. The data are acquired then processed and during the processing the data are d convoluted and D isotopes the file can then be converted and exported as LCS or as an MZ XML file and these files are compatible with the sim glycan software that can help automate the tasks of tandem ms structural characterization and I would just like to point out that the unified glycan workflow does support both to a B and rappy floor label glycans. It's important to note that the workflow that we recommend is a dual detector method and we do utilize both the flr and mass spec channels we use DSLR for glycan library searching and for quantitation and our quantitation.
Typically we use a normalization scheme where the flr peak area of a particular glycan is expressed as a percentage of the total peak area for all of the glycans identified by the library. Search we don't use the MS channel for quantitation we use that information to confirm our tu scientific library searches and I'd like to point out that our FMS has greatly helped out with our GU. With accurate mass workflow particularly for the lower buttons glycans and in this slide. I'm comparing the flr. Trace on the top to the mass spec base become a deterrent base peak intensity trace on the bottom and this shows really how closely the two detectors mirror each other and a particular interest are those glycans that later in the run and our president very low abundances as we can see in the inset using other tags we've really struggled to get high quality mass spec data to confirm the library search results but with our FMS. We're able to get enough of mass spec signal to be able to really confidently identify what those particular glycans are and as a general rule for me if I see a peak and the flr with an intensity of around 0.2. I'm quite confident that I'll get a good enough mass spec data to get that accurate mass confirmation of course this is a very rough estimate as all F all our detectors and bulbs are unique as our mass spectrometers. But again for me. This general rule works out quite well and so now. Let's talk about the our finesse. GU scientific library in a little bit more detail. This library is very similar to our to a bee library. So if you're comfortable using the - a bee library using the RF MMS library isn't going to be any problem the RF MF library was generated using nine different glycoproteins from several different biological sources. We use several monoclonal antibodies like etanercept infliximab trastuzumab and EPO. The nist monoclonal antibody pulled human IgG mouse IgG bovine. Fecha one for a source of highly branched in Sayulita glycans and yeast invert aids as a source for high mono species.
This library contains GU values for 177 glycans that we tailored to the biopharmaceutical industry and we realize this isn't an exhaustive library and some users may have some unique lichens that are in the library but unify allows users to create their own custom libraries that they can co search with the RFM s glycan GU scientific library. The library does contain all the major types of end link lichens it has high mannose glycans complex lichens and hybrid glycans. There's a wide variety of Sai allele ADA glycans including those that have n glycol ilnur Minik acid and and a settlor Minnick acid. There's also several glycans that have alpha linkle. Actos pairs and the library also contains GU values for many isomeric lichens including positional isomers and linkage isomers. So for example for several plugins we had we can distinguish alpha 2 3 links. Ayala Casas from alpha to 3 F from alpha to six links sciatic acids the development of the RF misc like in G scientific library with a joint project between Waters and Pauline Rudd and mark Hillard of the National Institute for bio processing research and training or nyebern Ireland and they're the ones that really didn't an exhaustive characterization of the glycans using a battery of EXO Casa de this for each bike and sample and using the EXO Casa de zzz's they were able to selectively remove specific monosaccharides and this gives us a very precise structural characterization. We gave a lot of information out of this. And this characterization is shown in the upper left-hand upper left hand side of this slide so for example they start off by running den digested sample and then they would first remove alpha to three linked silac acids and they could look at the shift in retention time in GU values and then they would remove the alpha to six linked silac acids and then they would remove you know subsequent monosaccharides so really doing a very thorough job of really pinning down what the exact structure of each glycan is and then each glycan has also been subjected to mass confirmation.
And for a lot of glycans. I've also collected mmmmm SMS data. And and. So yeah i'm very confident that all of the GU values and for the structures you know they are in fact correct. And why did we go through all of this effort of EXO class today's digestion z' and getting accurate mass data. It was really so that the users didn't have to do this. That all this work has already been done for you so you can just you know. Run your samples do the library search and get your results and I'd like to point out that each sample was analyzed at least ten times with the mass confirmation so we do have great confidence in the accuracy of the GU values since we are searching a library based on calibrated retention times. We do have a set of conditions for the separation that we do need to follow fairly closely to get the best library search results we do need to use the 2.1 by 150 millimeter acuity u PLC glycan b hm and column that has the 1.7 micron particles with a pore size of 130 angstroms. This is the same column. That's been used for the 2a b library so that part is the same the mobile phases mobile phase a is 50. Mm hola ammonium formate at ph 4.5 and mobile phase B is pure acid or nitrile with no modifier we conduct the separations for the RF MS experiments at 60 degrees C. It's bit higher than what we use for the 2a b library and we use a gradient from 75% b to 54% B over 35 minutes or a slope of zero point six percent per minute this gradient is a little bit different than the one we used for the two AP library but the column is the same. The mobile phases are the same just a couple of slight changes in the conditions from going from the to a/b experiments to the rapid flora math experiments. We use glucose units instead of raw retention times we found a GU values to be a little bit more stable across different instrument systems or across different columns or even across different days while retention times do tend to shift around a little bit and so if we search a library just based on retention times our search results may not be as well as we would like them to be so we recommend using GU values to be able to convert retention times into glucose units.
We first run an analysis using dextran has been labeled with rappy for rms and we can see this in the figure on the upper right hand part of the slide we then tell the software the exact retention time for the structures for example. We're telling the software that gu4 has a retention time of seven point. Five three minutes gu5 is ten point six six minutes and on and on and on we can then convert those GU values and retention times into a calibration curve shown just below the flr chromatogram and then we can generate an equation based on this data that we can then use to convert through tension times of our unknown samples into GU values which we can then use to search the scientific library in order to obtain the best GU values and get the most accurate scientific library search results for our FMS we do recommend using our rappy flora mess texture and calibration ladder it took a little bit of chemistry to be able to get our FMS attached to it but we were able to do that through an ethanolamine. Oh you're a linkage. And our dexter and ladder covers gu values from two out to thirty alignments to get accurate. G values from very small glycans out to very large ones. So you know some of these huge glycans that are you know highly branched with you know tetran ten ireri four silac acids of yukos's maybe throw in some like. Tosa mean extensions we can get accurate GU values for all of those types of glycans and to get the best library search results. We recommend running the calibration ladder with each analysis set and now let's take a look at some fully processed results that we get from unify and this slide is a recreation of what we would see on the left hand side.
We see the flr trace. That's been annotated automatically with the glycans from the library search result and if we click on any one of those structures we can pull out the extracted 2-iron chromatogram and here. We see two peaks when we do that so right away. I know that we're dealing with isomers. And if we want we can pull out the process mass spectra as well on the right hand side of the slide. We see the library search results and I'd like to explain how the library search works. We do search within a GU window we recommend plus or minus 0.2 GU units and after unify has calculated the GU value for a given peak in the flr chromatogram. It'll search the library for all of the glycans that fall within that window so for the highlighted peak in the flr trace unified determinate had a GU value of eight point one eight so the library search pulled out all of the glycans that had GU values between seven point. Nine eight and eight point three eight then. The software goes back and we'll determine which like and out of that pot out of that list is actually the correct one and this is where the mass spec data comes into play. It'll go back into the mass spec data it'll look for an ion and it can search different charge states. That's not an issue but it looked for an ion corresponding to that particular glycan in the mass spectra if it sees that ion with an acceptable acceptable mass accuracy about 2. Ppm then we say is that glycan is mass confirmed and if we look at the results for this particular search the glycan that's highlighted had a delta GU of 0.01 one GU units and a Delta M over Z of 0.005. So these are both very accurate results and so there's absolutely no question that the library has given us the correct glycan for that particular. GU value and now. Let's take a look at how we can use this workflow in a real-world study where we were comparing innovator and biosimilar inputs and map and can looking at the glycan profiles in flip.
The map has been on the market for about twenty years. Now it's an anti TNF alpha monoclonal antibody. It was originally approved to treat rheumatoid arthritis and it's now been approved to treat. Crohn's disease and a number of other diseases as well. There's been a biosimilar version known as in fly code has been developed by a company based out of Korea the biosimilar version has been approved for use in Canada Australia and South Korea and importantly this is the first biosimilar monoclonal antibody. That's been approved for use in the United States. And so we thought this would be a very timely comparison to do between the biosimilar and the innovator infliximab to begin our comparison of innovator and biosimilar infliximab we first looked at the flr profiles. We were able to obtain three samples of the innovator molecule in one sample of the biosimilar but if we look at the flr traces we can see how similar these four traces are. There are some slight differences with some of the lower buttons glycans. It appears but for the most part the major structures appear to be very similar we were able to identify 23. Unique lichens found the innovative molecule and 21 on the biosimilar. All of the glycans are on the biosimilar. Were also on the innovator molecule so there are two unique glycans on the innovator monoclonal antibody once we've done the library searching we've determined what the carbohydrates on the sample are we can start to group those into various classes based on their structural features. And this is what this slide is showing and we're just comparing different classes of glycans between the innovator molecules which are shown in blue and the biosimilar which is shown in red so we're able to classify in this example sugars that are neutral mono sila violated nazi. Consulate's be conflated mono and generate binary high menos and those that have alpha and K lactose units and I think overall we can see that the different classes are very similar.
There appear to be some slight differences for example it looks like the neutral glycans are slightly more elevated in the innovator sample and perhaps a mono thiolated are slightly more abundant in the biosimilar sample but overall we think that the the different classes are present at a very very similar abundance levels on both on both types of molecule to drill down into the data and really look at specific glycans that are different in their abundance levels. Between the innovator in the biosimilar molecule we can start to look at these mirror image plots and in this prodding at the biosimilar shown on top in green and the innovator shown on the bottom in black and the difference in the are on density of these peaks is shown as the red trace in the middle. Now we are looking at raw data in this example but we believe that since we start with the same amount of starting material 15 micrograms of each monoclonal antibody and we inject the same amount. We've already taken care of there on those agents. We believe that this is a valid comparison of things. If we look at speak at the retention time of 12.94 minutes we need to rotate it. Dips down then jumps back up which is indicating that there are no differences in the peak intensity for that particular models for that particular carbohydrate if we zoom in on some of the carbohydrates that are lower their abundance levels absolute later around 18 minutes to about 24 minutes. We can start to see. There are some some differences those abundance levels the interesting thing is that these are all terminated within glycol in Germanic absolute and they all appear to be more elevated and their abundance bubbles on the biosimilar monoclonal antibody although it is worth noting that these are present at very low abundance levels when we start to look at all of the innovator monoclonal antibodies we can start to see those patterns those on the mirror image.
Plots starting to really be emphasized for the glycans that are capped with and glycol aromatic acid and overall looks like that those types of glycans are more abundant on the biosimilar molecule. We have four examples of those in general. It looks like these dilated species were three times more abundant on the biosimilar molecule as on the innovator molecule if we look at glycans that have alpha linked galactose pairs. Those tend to be more prevalent. On the innovator molecule in fact this hybrid glycan shown on the bottom right hand corner was only observed on the innovator molecule an interesting lichen was the one that had both an acetylene or minik acid and alpha-linolenic. What i find interesting about this was it was more abundant on the on the biosimilar molecule although it wasn't as abundant as some of those some of the other carbohydrates found that didn't have the alpha galactose linked to it so. I think it's really interesting when we look at these glycans that can have some immunogenic responses that there seems to be a preferential expression for the for the type of map again the the glycans that have n glycol aromatic acid tend to be more abundant on the biosimilar. The glycans that have alpha like galactose pairs appear to be more abundant. On the innovator molecule it's also important to take a look at the levels of yukos and again we can start to look at that those levels once we've done the library search over all the levels of fucose were quite similar on both molecules there was one particular glycan that we call. FA - six G for one that appeared to be more abundant on the innovator monoclonal antibody but that increase was kind of compensated for by the s6 a tube-like and so overall the levels of the constellation on both types of map were very similar we concluded I'd like to highlight some of the key points that we discussed. We're very excited about right before. M s this is the first molecule on the market that can rapidly or instantly label.
What console amines that can enhance both the flr and the mass spec signal again. This reaction is done in a matter of seconds but we recommend letting it go for a few minutes. We can then integrate this into a simple sample preparation workflow. We've been able to cut the time investment down from several hours to a few days down to just a few minutes so we can. We can get the sample preparation done in as little as 30 minutes. But depending on the number of samples it may take up to an hour we can then incorporate this simplified sample preparation scheme into a simplified data acquisition and analysis within unify. We're very excited about being able to simplify the data analysis using our scientific library search glycan values traditionally. This has been an area. That's been very difficult and can be very time-consuming but with the ability to search a scientific library greatly simplifies this task and it makes it more straightforward and a lot faster we can then take this overall workflow of FRA performance labeling with our automated data acquisition and library searching we can start to develop some very interesting applications and in the example that I showed today we can start to compare monoclonal. Antibodies from innovator and biosimilar companies and we're able to look at what types of glycans are different in what types of glycans are the same. And so with that there's a and with that there are several people that I'd like to thank I would like to thank Pauline Rudd and market leaders. Nyberg they've been instrumental in assisting with the core development of the RF. Ms like in geo scientific library. I'd also like to thank Matthew Lauber Yin Jing to ROS Waller Scott Berger MN do Jennifer Fournier and Alexei Armstrong. The waters for helping put this presentation together. I'd like to take this opportunity to thank Jeff for giving us this opportunity to present some of the research that we're doing up waters in the area of glycan analysis.
We're very excited about some of these developments that we believe will really streamline the task of glycan analysis. Thanks bill that was a great presentation you provided audience with some great information about the advantages of the rapid for MS kit and how it will streamline their glycan analysis workflows so we thank you for that before we get started on the Q&A session. I want to remind everyone once again that this is your final chance to submit questions for our speaker so hurry up and submit them now just type them. In to the box to the left-hand side of your screen and hit submit all right to have some really great questions that have already started to roll in. So why don't we get to answering as many of them as we can bear with us for a few moments while we get everything situated on our end and we'll begin the Q&A all right everyone thanks for joining us for the Q&A have a few questions that have rolled in already. So let's get to them. Bill first question is. Can we add glycans to the library. In addition to the more glycans. I have some very unique structures and I want to be sure that I can use the library as effectively as possible. Yeah that's a great question. Unfortunately we can't edit glycans that are in the library. We've got that locked down but what we can do within unify is we can create our own custom library so for the customers that are dealing with some unique structures that they've engineered for their particular protein. We can go in. We can create that like lichen structure. We can calculate the data and we can create a unique library. Put that glycan in that library and then we can go search with the with the rms geo glycan library and we'll be able to get the structure identified that way but but we can't edit the actual library itself and go searching. The library works just fine. Works great alright so our next question one of our audience members would like to know. Can we use the library outside of unifying.
I'm using our for glide analysis and would like to be able to use this library. I yes you can. This was co-developed with with knivert. And this has all been uploaded into their glyco base database. So you can get this exact same data through glycol base. They have to register for an account with that. But you can search the data it is manual. It's not automated like what we would do and unify but you do have access to that data outside of unify. Yes okay great. Another question audience member starts off by saying. I'm working with some very large glycans that elute very late in the lc/ms analysis. How much flick is there with the PSU PLC condition. We do have a little bit of flexibility particularly with what the length of the gradient so we run a gradient from 75 percent zero nitrile to 54 percent of cedar nitrile if you need to drop that down to say 45 percent or 40 percent that's fine that's not going to affect anything you can you can lengthen the gradient but we do need to keep the gradient at the 0.6% per minute what we can't do is change the temperature. If we start to change the temperature then we start to change the selectivity of the column and then our library searches aren't artists as as accurate as we'd like them to be but lengthening the gradient not a problem if you want to shorten it a little bit. That's not going to be a problem either. Just make sure that we keep the slope at 0.6% per minute and and but we'll get very accurate. Library searches. Thank You boom another question. Will the decode will the deck like oscillation step eventually give a monosaccharide or was it a mixture of mono and oligosaccharides the dig like. Oscillation step we're just using the rapid beam jeiza from New England Biolabs and it you know it's the access as normal P&G ASF does and it will release intact and link likens as glycosyl amines. So yeah we don't release those as monosaccharides. Alright thank you we have one person who will show. Kaiba and I just wanted to mention the one thing that we we have done is we have optimized the buffer with the rapid P and JSF so that it does work better with rappy floor ms so the buffer is a little bit different.
That's why we're cool marketing with with. New England Biolabs. Alrighty thank you so water. Navarro I guess missed a little bit information and would like to know if you could repeat. How do you maintain the glycosyl mean during PNG ACF treatment and what keeps it from converting to reducing it. Yep that's a great question so the ability to to do that the D glycosylation is we do use quite a bit of high concentration of the P and GS F and we do the reaction in five minutes. And the the half-life of the glycol amine is a couple of hours so by getting the digestion done in five minutes and then doing all the subsequent sample of derivatives ations. You know in another five minutes or so. You know we're able to get the we're able to traffic lichens as glycosylated efficiently. Before they have a chance to be converted back down to the to the hydroxyl so really we're just taking advantage of the rapid kinetics and being able to wrap it too rapidly label things so we're just taking advantage of the time okay and it looks like we have our last question. Audience member would like to know. How do you determine the stereochemistry of glycosidic linkages it based solely on the glycan library. Yes so this is all ties back into the very rigorous characterization with the expo galactosidase digestion that. There are co-workers and collaborators that. Nyberg did or they're using very specific XO glycolysis to be able to you know very selectively remove amino acids based on you know alpha beta linkages. And you know alpha to 3 vs alpha to six linked sciatic acids and you know the the galactose linkages as well and and on and on and on and on down for each each monosaccharides so we're getting that type of information from our very rigorous characterization using very specific XO qui casa dayses alrighty.
Well thank you bill and with that. It looks like we've come to the end of our webinar so want to remind everyone that our webinar will be archived for up to a year on our website at WWE and it you can watch it again or feel free to for the link to your friends and colleagues which we always recommend. I'd like to thank Bill again for his informative presentation and I like to thank the audience for their attention and thoughtful questions and the very special thanks to Waters for sponsoring this webinar. So hopefully we'll see you again at another gen webinar in the near future. Good bye for now.