Webinar: Innovative Methods Advancing Discovery and Characterization of Antibodies

0:00 Cheri Salazar: Hello everyone on behalf of Cartera and Biointron I would like to welcome you to our webinar Innovative Methods Advancing Discovery and Characterization of Antibodies my name is Cheri Salazar and I'm the host and moderator for this event if you have any questions you would like to ask our speakers please go to the bottom of your screen and click on the Q&A button then type in your question we will get to the questions after the presentations i'd now like to introduce our speakers dr Long Xu is the R&D director at Biointron and Dr Daniel Bedinger is the senior manager of field application science at Cartera we will start the webinar with Dr chu dr chu please go ahead

0:52 Long Xu: Thank you Cheri for your introduction uh good evening everyone it's my great honor to have the chance to present here today I will talk about the new technologies and platforms to empower antibody discovery and over the past three decades the development of antibody drugs uh is promising and rapid we are seeing more and more antibody drugs entering the market to benefit human health 2021 saw FDA approved 100 antibody drug so now we have entered the error of 100 plus antibody drugs and the trend looks very encouraging but the antibody drug

1:38

industry has been uh troubled by a number of bottlenecks such as uh big size po tissue uh penetration and difficulty in modular uh coupling and immunogenicity problems no affinity and and so These bottlenecks call for new technologies and platforms to break through and we believe single domain antibodies offer new options in antibody drug discovery as we all know in some cases and shocks uh there are some heavy chain only uh antibody with light chain

2:17

missing the variable domains of these heavy chain antibodies are often called VH or narrow bodies they have a lot of unique advantages such as uh small size good tissue penetration high stability and great potential as modules for downstream engineering good solubility and high homology and flexible production system they're so good and then how should we use VH okay here just to show a a simple workflow and usually we start with Alpaca immunization to generate the animal library and then we screen hits from the library and then a lot of uh functional essays uh performed to identify uh lead antibodies um with good

3:16

potentials then these uh lead molecules are subjected to optimization like humanization and affinity maturations after that they are ready for the uh downstream applications say the construction of five specific antibodies there are many uh methods for the antibody screen and today I just talk about phase display and single bell sorting platforms at by Intron we have abundant experience in VH discovery we have our own Araco range with over 300

3:58

Apacles and this picture demonstrates uh the phase display workflow and the starting point is immunization of an ARPA and then PBMC is uh isolated and then the VH genes got amplified and inserted into Fate M and then transformed into competent cells for the construction of VH libraries and uh which enable uh a vast connection of VH variants get displayed on phase surfaces and the phage library usually have a pretty big capacity uh say usually more than 10 to the 9th power okay then uh how should we screen hits usually we perform several rounds

4:58

of biopenning and after each panning the positive phase that can specifically bind to the antigen and get isolated and amplified usually after three rounds of panning the positive rate uh in the library can achieve double digit percentages at this point we can pick up monoconal phages from the library and then perform elyses to identify good positive phages and then send them uh for sequencing to to obtain the VH sequences next I will use example to show the process and um in this case the antigen is called X it's a receptor expressed on the surface of T- cells okay the um apical immunizations and construction of fage library is pretty straightforward i just

6:01

uh go directly to the panning uh stages and here we use both solid phase panning and liquid phase panning and in solid phase penning and after three rounds of penning actually we uh picked over a thousand clones and actually we uh find out that uh 360 positive clones which result in 11 unique sequences and in liquid panning the results are similar and we finally uh got uh 900 something positive clones and which resulted in 43 unique sequences so totally we obtained 54 unique sequences as our hits okay so the process usually take uh more than four months and many people may say okay that uh phage display approach is good but it's still very

7:03

time consuming can we um have a more powerful platform and cut the time by half actually by intro we have the new single B cell platform called AB job yeah it can and work as a ultra fast antibody sorting platform and and and accomplish the the the and get the job done uh uh we reduce the uh you know time we reduce the timeline okay and uh actually this just to show okay the conventional uh work process here actually many steps can be replaced by uh by by this uh AB job uh uh system which can got you know this tedious job get done within one day so the total totally the the the uh

8:03

timeline can be greatly reduced by two to three months which is significant and this slide just to show how the platform work and actually it can uh generate tiny volume uh oil or droplet and each droplet can can contain one uh BCL together with antigens and then pass through the detector one by one the detector can tell us if the oil droplet contains a good B cell or not if they're good these good B cells got pulled together and then sent for sequencing to obtain the the candidates uh sequences this just one example and uh uh after one example after one experiment and on the AB job platform we can uh obtain

9:01

sequences okay and then we uh can uh we use carterra to perform the affinity test to uh to to to identify the affinity and yeah and in actually this case uh we are doing the NTPD1 uh antibody discovery and the results is pretty good we uh obtained 62 heat and uh uh the affinity test turned out that most of them show very good binding and we all know that there are many uh PD1 antibody drugs uh on the market and we are curious if our heat shows the similar uh binding epitope with any of these NDP1 antibody drugs so we performed the epitope binding on Carterra it is very powerful uh

10:00

application actually the the uh say if we have say uh 50 heat the Carterra uh system can perform the 50 by 50 competitive binding experiment and then the the 50 heat can be categorized uh based on the similarity of the binding epitope okay so here's our results and we performed the binding um with two uh b similar and uh the the results is pretty interesting we found out that 19 heats uh show very similar epitopes uh as one of the b similars but none of our heats have the same epitopes as the other one okay um okay i have finished the the the the introduction of uh uh face display and single B cell platforms and then

11:06

next step is uh the functional essays on heat to identify the good lead antibodies and we begin with Eliza essays and okay the binding are pretty good we also uh performed the SPR test on uh Biacore 8K uh okay the PR results are also pretty good okay so you see we um use both Elisa and Biacore to uh perform the affinity taste they both have advantages and disadvantages say Eliza are cheap and convenient and it can work with uh supernatants and biacore can provide very accurate uh affinity datas but but the the the instrument is

12:00

very expensive and the operation is complicated and also they have uh higher uh requirement for samples and also uh the vehicle is not high throughput thus we we want to you know have a good have a new option that can combine the advantages of both and fortunate fortunately we have the Carterra uh can you know make up for the those shortcomings and Carterra is high-throughput and it can work with superance it can provide a good uh k on kds and um the the the data analysis case is very powerful and broad so carterra is widely used at by intro in antibody discovery and characterization and here I just show and the cater can be used at every stage of uh antibody discovery and um characterization here uh the car is used to uh perform affinity test uh you know

13:12

over uh hundreds of antibody hits at a time which can save a lot of time okay besides affinity we also uh want to know the uh the the the um broking ability of our kids and here we performed the ligand blocking essays and the results showed that 44 have clearly uh ability to block the interaction between the ligand and the antigen X which is pretty good okay and uh so and after these uh functional essays then uh we identified several uh lead molecules but we know that most of VH have low affinity so affinity maturation is a must at Bon we

14:05

have the FC mass uh platform which called the full coverage Mammalian expression system for affinity maturation and it uses a side saturated mutagenesis in each site in CDR regions to introduce the uh sufficient mutation uh diversity and uh basically okay for each site in CDR regions the the original amino acid is replaced by 11 other uh uh 17 other amino acids to generate a side saturated mutant plasmid library and then these uh site saturation mutant plasma library uh transformed into mammalian cells amplified and extracted and then transect into mammalian cells for

15:01

antibody expression and usually we end up with thousands of uh supernatants okay expressed by thousands of single point mutant antibodies with these supernatans we perform the high throughput Elisa assays uh to uh screen for mutation hotspots hotspots refer to the single point mutation that can significantly improve the binding ability compared to the parentals okay then uh we can obtain a lot of useful mutation hotspot information uh we select the goods hotspots and do the combination to design dozens of candidate antibody sequences which are expressed and uh characterized by uh Carterra and some other uh tests okay here just just examples you

16:03

know we know that say some um some uh VH they they they k improvement but say the right one the the k off is too high okay so we we we we can uh improve their affinity and special and um uh respectively say in this case we significantly improve the k on and can and result in uh more than uh 100 fold increase in uh affinity and in this case koff is too high with just four hotspots combination we achieve and say as high as about 23fold increase in affinity okay so after this optimization and then the VH the the sequences of the

17:04

uh lead molecules are ready for uh the final applications and by intron we are very experienced in construction of a variety of bispecific antibody formats uh okay so I will just give a quick introduction to B in we have centers in China and in the United States and the Europe and in Europe and we are space uh uh we are focused on the antibody production discovery and optimizations and we have state-of-the-art facilities and also uh build very uh advanced automated uh workstation which can enable us to provide very reliable and faster deliveries and uh so that's all of my presentations thank you and now I turn it over to Dr Bedinger

18:03 Dan Bedinger: well thank you Dr Xu let me share my screen here okay i included me all right well thank you Dr shu uh so today I'm going to talk about some new approaches and applications we have for the LSA and Carter's ultra platforms including how these approaches can apply to multi-specific antibodies so I'd like to start by introducing the Carterra ecosystem as I like to call it um we now have three instrument platforms our original LSA

18:49

the sort of updated more sensitive LSA XT and our new ultra platform which has an emphasis on small molecule binding applications um and just general higher sensitivity we also have a software environment that includes three software packages the navigator software that we use to design experiments and collect data um then there is two analysis software packages that you know they all come together as a bundle um one is called kinetics and one is called epitope and it's relatively self-explanatory what those focus on and then we have a whole line of biosensor chips and consumables that you can use to run these assays so the LSA platforms and the Ultra are array based SPR systems and these are really ideally suited for a lot of the core applications in biologics and antibody drug discovery being kinetics and affinity analysis uh competition based epitope binning uh and peptide mutant mapping and quantitation and in addition to having a hardware architecture that makes all of these assay formats both very high throughput and not consumptive of sample uh we put

20:03

a huge amount of effort into designing analytical software packages that make processing all of this data fast and rapid but also visually rich including uh features like in our epitope analysis software we have these heat map and network plots that can present a lot of information in a concise way so in this talk we're going to talk I'm going to talk about uh the LSA XT and looking at FAB kinetic analysis from crude pariplasmic extracts this is where we can measure the full on rate off rate and affinity of these without the need for purification we'll talk a little bit about epitope binning and its role in bi-specific or biparatopic antibody discovery and then also talk about some assay formats that are more specific uh for bi specifics including independence or non-inference of each binding site and bridging assays and then I will also introduce the cartera ultra which is our new platform and enables the binding of

21:05

screening of small molecules and fragment libraries to many targets simultaneously so the LSA and Ultra platforms both use a a similar architecture shown here where we have two semi-independent fluidic systems the 96 channel side is used to flow 96 samples at a time on the surface this is used typically for array preparation or capture and we can do in the LSA four prints of 96 to create a 384 array or on the Ultra it's two prints and 192 and then the 96 channel mechanism can move out of the way and a single flow cell will dock over that entire array area and do serial injections um either of different concentrations of antigen for kinetics or of competitors for an epitope pinning

22:00

this is what we call one-on-many analysis and it's what gives uh these platforms their really high throughput and low sample consumption and talk about this in a little bit more detail um this is an animation of the 96 channel side of the instrument so you can see that the 96 needles will descend down into one of three micro plate locations and draw the sample to the chip surface and flow it back and forth so if this is a capture where we have an affinity based surface um we can flow those samples back and forth over the surface for a long period of time you know 40 minutes is not atypical um and get a good enrichment of the antibodies or whatever the lians are in the samples to the chip surface um this is allowing for long contact times under high flow rate without having to uh increase the sample volume or decrease the the flow rate again so I'll run this video again real quick um if you do this you know once for 96 then you can go wash the needles get additional samples and um do up to four

23:05

of these on an LSA to create a 384 spot array which is shown um in the last image here when the flow cell and docks it's a higher density array where the injection or the spots are interlaced then the single flow cell can dock over that entire area where the array is printed in flow one sample so again a one 270 microL sample volume gives you binding information across the entirety of the array this is what allows us to do this oneon- many analysis and gives the real efficiency so you know one injection per concentration on a kinetic titration um or one injection of a competitor in an epitopine will give you sandwiching information for an entire array of clones so we'll move right into an example here u we're going to about this is the capture of fabs out of parlasmic extracts which is a phage panning

24:05

workflow you can induce an expression of the fabs into the paraplasm and then create extracts from that um to do this we used a stripiden sensor chip and hc30m chip and captured the anti-CH1 so which is a fab constant region domain antibbody that we purchased from thermo it's actually a VH antibbody um so it is a nice monoconal for capture we inject that over the strep aident surface create a lawn of about 1,200,300 RUS at which point then we can flow the 96 well plates of fab samples so here you can see three captures um these are crude samples so relatively straight bacterial extracts so we get a large bulk refractive index shift during the flow and at the end when it returns to buffer there is a tiny amount of fab about 75 RUS captured onto the array these are

25:04

very low expressing samples and you need a long capture time to get enough enrichment to the chip surface to be able to see good kinetics this is actually a 40inut uh flow of those antibodies to capture them and but that is long enough to where most of them achieve similar levels of immobilization um and uh we can get good kinetics from that so at that point then the single flow cell will dock and we will do a series of blank injections to stabilize the surface and then an increasing concentration series of antigen and that is shown here where we're going from 1.9 nanomer to 500 nanomer in a four-fold series and from this we can fit the on rate the off rate uh the Rmax and the affinity using a one one langmir model so this is double reference data at this point and if we look at that in aggregate this is showing the 96 clones captured to the array in triplicate

26:05

um so this is a single run this entire analysis used 7.8 micrograms of antigen to generate all of this kinetic data because remember it's one injection of each concentration to generate all of this data in parallel um there's some highlighting here in the data so spots with very low signal are called non-binders and shaded gray things that are purple have a kinetic limitation where the antigen concentration wasn't high enough to drive towards saturation and give a well-escribed Rmax so that is flagged and then things that generally have high error in the fit are flagged in yellow so there's some automated uh QC profiling of the data in this but and if we zoom in on some of those so this are two different antibodies each captured in triplicate you can see we're able to get well resolved kinetic profiles with full kinetic information from these and on the XT this can be done at a at a relatively low signal

27:04

level these are less than 10 RU sensors for all of these and we're getting uh consistent kinetic estimates across the three replicates even at the lowest the third replica these are the third time these samples were actually captured onto the array and we started seeing a little depletion or dilution of the samples so signal is a little smaller but it's still giving us the same kinetic result so we were able to get triplicates from a single well of bacterial extract in that test okay moving on we'll talk a little bit about epitope binning so the Carterra platforms really enable pair-wise competition based epitope bin at a new scale um we do this by creating an array of antibodies so you can put up to 384 unique antibodies on an LSA array sort of 90 I guess on a um ultra and

28:02

then we use two assay formats one we call classical sandwich and that is where you inject antigen followed by an injection of sandwiching antibbody and then regenerate the surface and repeat for each clone so you generate a parise data that way if the antigen is multivalent a premix approach is used where you mix excess competitor antibody against a fixed concentration of antigen and inject that over and look for either sandwiching or a loss of that so after those runs are completed uh we open the data in the epitope software and we have some great visualization and tools to do this processing in there so shown here is a classical sandwich u data set this is from one spot so you can see there are many cycles injected over it so one antibbody in the array where we have an antigen injection the green bar will is normalization so it sets the binding level to 1.0 so even if there's some variance across the assay it scales everything

29:04

equivalently the second report point bar here is the report point or uh and that is actually the value that will populate the heat map table so you can see the blue traces are buffer control injections and so the value that is populated into the heat map is actually the difference from the proximal control so even if you get a significant amount of dissociation of the antigen from the captured antibbody because we're doing it rel the analysis is looking at the normalized signal relative to the controls um we're still able to d differentiate that often using a global cutoff so once uh these data are set with the the time points and the cutoffs the heat map is populated in this table the rows are immobilized lians or antibodies and the injected columns are the or the columns are the injected

30:02

analytes if it's green that means that these interactions are sandwiching if they're red they're blocking so we can create this pair-wise map of the way these interact and the software will automatically sort these to cluster um like bins together it will generate dendagramgrams to show how these relate and then construct these network plots in these network plots every clone is a node either a circle or a square if a line is connecting them that means that they are competitive with each other and there's a blocking relationship if they're contained within one of these colored uh regions we call these epitope bins or communities depending on sort of the level that you're analyzing the data at and they mean that those clones share a near identical or very similar uh binding profile in the heat map so looking at this a little more again the immobilized antibodies are the rows the injected antibodies are the

31:00

columns this is an example of a large single epitope binning experiment in this case there's 170 immobilized antibodies that were able to be included in the analysis and 188 analytes this is so about 32,000 interactions are shown and this is a a good performing assay that's pretty representative of the way you would like these to look we have the self versus self being highlighted in the black and they're all competitive with each other which is good and we have clear symmetric behavior where we're having similar patterns of competition in both orientations uh when you look at it and the software can cluster these and we can interpret these to a relatively fine uh resolution of epitope behavior so here's an example of a community where all of these clones are competitive with each other they share many blocking and many sandwiching relationships so this is a nice coherent community that is uh very descriptive of this epitope behavior of these clones here is a neighboring bin or

32:03

community and you see and see they're also competitive with each other but also competitive with the clones from the previous bin and they share many sandwiching and many blocking relationships but down here and over here there's a difference these are sandwiching with clones from this community whereas these are blocking so we've now included this to enough clones and enough diversity in this where we can do fi fairly fine resolution differentiation of shifted epitopes or you know so these are overlapping competitive epitopes but they are different in that they have sort of moved around the receptor and you can continue going through the data and categorizing these different activities you know here's another overlapping epitope bin that actually has a little bit more differentiation has additional blocking and then a different behavior versus subsets of these over here and you can kind of walk through the entire data set establishing these commun communities and these patterns of of

33:00

shared behavior to identify at a at a high resolution these epitope clusters um so how how can this be applied so here here's a published example from science translational medicine josh Tan's group at NIAD um did this they were trying to make therapeutics to SARS CO2 and they identified it a number of potent antibodies where they did affinity characterization on the LSA and and kinetics and epitope binning and they found that they had some potent neutralizers from a number of bins and in and they wanted to try to make cocktails that may have synergistic potency by neutralizing different epitopes when they did that combination they found that really the potency advantage was additive there wasn't a clear synergy so they say "Okay well maybe we can make a biparatopic uh using the DVD IG strategy which was uh one of the formats included on Dr shu slide." And that so they they made a combinatorial matrix of a number of

34:06

potent neutralizers from different epitope bins and characterized them in this DVD IG format uh here's an interesting example of a demonstration of the actual bioparatopic nature of these clones so these are classical sandwich bin assay format using two different lians so these are the monospecific 503 and 664 we will call them uh we get antigen binding if you inject the same clone it does not sandwich if you inject either the 664 monospecific or the 503664 by paratopic we do get sandwiching and the same in the inverse so if you have 664 immobilized and injected it blocks if you inject 503 it sandwiches and if you inject 503 664 with either of two different linkers you get a good sandwiching so this is confirmation that these biparatopics uh the 503664 actually u recognizes both distinct epitopes uh within that and then these were tested for potency and there were examples uh from the combinatorial set that showed a 100fold increase in in synergistic

35:20

potency versus the combination of two clones alone this was looked into structurally and it was actually a really interesting behavior where two different monomers within the RBD spike were being sort of cross- linked together forcing the entire spike molecule spike trimer into a totally

35:38

inactive confirmation so rather than just being a sort of stokometric blockade of binding it was causing the entire receptor um to be sort of knocked out with a biparatopic molecule binding it this is sort of exactly I think what they were hoping to see we'll talk now about a few sort of bias specifics this is assuming your pro your uh binder is actually binding two different targets at say formats to look at so the most basic characterization

36:09

would be the affinity of your topic constructs probably comparing them to the parental you know single specific controls so this is basically just standard capture kinetics for us but because the LSA does these things in high throughput it's actually very easy to test large panels of these so in in this example we'll say we have uh 12 mono specifics to each target you can construct a library of 144 uh commentorial mutants of those and you can run those in a single afternoon or evening affinity this can actually be expanded on the LSA up to about 1,152 unique lians in about a two-day run and you can do you know multiple analytes to those so this would be an example of data from that so if we have uh eye

37:02

specifics targeting you know antigen a this is the these are the parental controls and then in each row are the constructs made using that sequence you can see that generally we get good recapitulation of binding of those uh in the biic constructs there's a few examples uh where maybe the binding is impaired and then if we look at the antigen binding in the same array um you know can be the same capture kinetics experiment we get the 12 examples of the bipartopic in this case all of the antigen binding is retained in these constructs with equivalent affinity so this would be the desired result of that as well so there's some other tests you can do on bicep specifics um you may want to know if I'm occupying one site do I affect the binding of the alternative antigen Um in this case you can if it's one of

38:01

them is high affinity you can just saturate that site with antigen uh and then inject a titration series of the other and get full affinity if say antigen B interaction is weaker and you're still trying to measure the antigen A you would need to do several concentrations of antigen A in the presence of a saturating concentration of antigen B sort of include injections and reference that out so a little messier but still totally doable and you can clearly identify uh the binding profile of A we have an example of that this was a high affinity B binder so the in this case the B is unoccupied and we get uh kinetics and then in this example the B site is occupied and the kinetics look uh largely similar very much the same so in this case uh occupancy of B does not affect the binding of A and then the last bicep specific uh format is one that truly demonstrates that you can have occupancy of the molecule and this is probably the only one that really

39:10

totally definitively shows that so this would be you could say coat the array surface with or capture you know antigen A and then use the 96 channel mode to flow your biopec so they will interact with the surface and attach via their antigen B specificity and then come in with injections of antigen A and you will only see antigen A binding to that surface um if you know they're binding via the antigen B uh specificity which is capturing the antibodies to the surface so you know this can be done at scale uh of antibodies automated to profile uh and and demonstrate the dual occupancy for a large number of clones in a single experiment okay so now I'm going to talk a little bit about our newest platform that we're excited to introduce which is the LSA

40:03

Ultra so the LSA platform was our first platform and was really well suited for antibody characterization and other protein uh protein interaction analysis protein library screening the LSAXT works just like an LSA it just has higher sensitivity um this really brought us down into the range of Protac and other sort of larger small molecules in the 500 Dalton range um and the ultra um takes us all the way in sort of therapeutic molecular space or at least investigational with fragment binding you know down to the 100 Dalton uh range the the difference in these platforms are all detailed in this table um if you look over here at the Ultra we're getting a lower signal to noise than the LSA or XT um the data collection rate is faster we've actually have less spots on the array there's a bit of a trade-off there to increase the

41:02

data collection rate and the decrease the noise um but we now have sensitivity as I said down to sub 100 Dalton or 100 D alton and the the single flow cell fluidics module is almost entirely new where we can now rapidly process injections down to 180 microL sample volume where the LSA is 270 which opens up the sort of standard 200 microL depot plate format and the increased speed and sample processing allows you to do 384 single flow cell injections within a day um and that can include things like 50% DMSO washing of the single channel fluidics it also has an expanded thermal range where we can now do full kinetic analysis from 10 to 40° instead of 15 to 40 like the LSA and LSAXT um so the enhanced sensitivity is demonstrated here this is a injection or

42:02

a series of kinetic analysis looking at the binding of ethane sulfonomide which is a 109 dalton fragment to carbonic and hydrates which was cap captured to a nickel surface these are actually duplicate injections of each concentration shown at uh many spots within the array so you can see we get um very low noise very reproducible binding even of uh sort of challenging small molecules so what what can we do with this you know definitely small molecules hasn't conventionally been something where people are looking for high throughput array based thing and we think that there's some opportunities to assess specificity and binding selectivity in a different way with this platform so here here's an example study where the Maybridge 1000 fragment library was purchased and screened against a panel of 125 biotinilated kinases provided to

43:00

us by carabio this allows you to select positive binders out of this library against all 125 kinases and then you can run follow-up experiments to look at affinity or lian deficiency determination so when you run this uh 125 kinases against 1,000 fragments this is a 3-day experiment you end up with 125,000 interactions of data points um this experiment was done at 300 microar compound concentration in a fairly complex buffer with glycerol magnesium chloride uh DMSO and some DTT which is required for fragment solubility and u the kinise activity in the array this was done at 15° C and in this animation that's playing over here on the right these are the actual fragments from the Maybridge 1000 library being shown at the effective screening rate of the platform so if you consider 125 kinase and an injection of a fragment every 3.5 minutes you're actually getting an

44:05

effective data point about every 1.7 seconds so this is really transformative we think in this idea of fragment screening where you could do entire families of proteins against the library and understand selectivity upfront be great for you know s applications and things like that um we have a newly developed software that helps process these huge data sets again this is about 125,000 interactions we're getting a thousand compounds against 125 receptors so we have these uh report point plots where in this case we were using ADP as the binding control for the kinases every um 10 cycles or so and that allows you to track the activity of the kinases over time and we're able to see uh the binding positive fragments and call them out against all of these targets simultaneously so we're talking about

45:02

things like families of proteins like kinases if you have inhibitors you can run concentration series and get kinetics of these to the entire panel of kinases simultaneously and so in a single experiment you can do things like populate a kinome tree where here we have the affinity of the interaction shown as the size of the the circle on the different kinases are represented and we've done some of these where we can compare the affinity value to the uh neutralization that's done from the cell-based assays that people typically uh pay you know third party CRO to generate this data and they match very closely so taking something that was quite an expensive and large endeer to get this kind of information being able to run it in a single experiment in a short amount of time is potentially really valuable um so uh with that I can wrap up um you

46:02

know HTSPR is an efficient way to generate high quality kinetic data for large panels of binders even from crude samples uh the LSA enables workflows to address special needs for binding characterization of bi-specific and biparatopic binders and then the carterra ultra is now enabling fragment and small molecule drug discovery screening you know it has high sensitivity with advanced microfitics optics and enhanced thermal performance that's allows to do this it's flexible it can handle nearly any sample type from fragments to antibodies um and the uh new advanced single flow cell architecture can reduce experimental run times and make the processing of these samples faster and we're working on and have you know next generation software tools that make processing this really high volume of uh small molecule screening data um tractable and

47:00

accessible so I I'd like to make some acknowledgements at Carterra um Rebecca Rich and Tony Gianetti organized and ran these fragment studies uh Tim Gur leads our commercial team uh the FAB binding data that I showed is actually a collaboration we had with Pfizer and Christopher Brandstrom in San Diego and obviously Josh Tan at the NIH for its excellent work and with that we can open it up to a Q&A i think  Cheri is going to Yes

47:35 Cheri Salazar: I'm here thank you to both speakers thank you Dr Bedinger for that and Dr Xu for uh a great presentation we will now answer as many questions as we have time for if we do not get to your question we will follow up with you after the webinar the first question is for Xu and this looks like it's a two-part question uh what is the sample size per experiment on the single B cell sorting platform and about how many positive cells can be obtained

48:15 Long Xu: thank you uh that's actually a good question to highlight the out how high throughput of our out uh platform uh typically we load 1 million B cells in each experiment and which uh usually give us say uh thousands of positive B cells

48:38 Cheri Salazar: okay thank you uh the next question will be for Dr Bedinger in epitope binning if different antibbody spots have different binding profiles are you always able to find a global cutoff value to describe blockers and sandwiches for all clones um no in a perfect world that would be the case um but really what you see when you do these especially large epitopines is really a broad range of behaviors you know some with very rapid dissociation some with uh you know very stable binding some that have a little bit more sort of sticky profiles where there's a little different behavior versus some of the lians so um while there the normalization helps a lot to make the

49:24

signals similar um they're not all going to be the same and so one thing that's actually quite powerful in the epitope binning tool is the ability to customize the cutoffs per lian so it it can be a bit of an effort to go through the data set and do the curation but it really pays off in the end where you can set custom thresholds for individual lians um and make sure that you're making the right calls in your binning map in terms of what is sandwiching and what is blocking so I think that's one of the really powerful aspects of the epitope binning tool is it allows you to really address a lot of the complexity that's inherent in these experiments

50:02

great thank you um it looks like this next question will go to Dr chu uh how many fold can the affinity of an antibody typically be increased through affinity maturation okay it really depends uh the answer varies it's a case by case okay so usually for uh um parental antibody with affinity in the narrower range or weaker uh we can see say 10fold or even uh higher increase in affinity but say if the parent anybody has affinity say the KD is less than one nanom it's more challenging and but we can still guarantee at least a five-fold increasing affinity

51:02

great thank you uh the next question uh this one would be for Dr bettinger if an array has 384 spots how can you screen up to 1,152 samples in a run

51:17 oh yeah good question and really talk about that so um if if you have a capture surface so an anti-FC an anti-H uh protein AG you know anti-V5 tag or whatever your uh system is that you're using to tether uh via affinity those ligans to the surface the surface is regeneratable so you can have 33 384 well plates in the instrument you know with high diversity of clones or or samples or even replicates within them um and you can capture that 38 84 at a time strip it with regeneration uh if you're using nickel NTA or something you have to recharge the surface and then you can load additional samples on so that can all happen within a single run we can have many sort of we sets we call

52:07

them of analytes or kinetics in those if you are doing something like epitope binning where you're covealently linking the antibodies to the surface then you are sort of limited to the 384 spots because they don't go anywhere they stay attached great um we have time for one other question uh during the live webinar uh and this one will be for Dr Bedinger um how do you measure the affinity of a biparattopic binder that's a good question right because with a bispecific you you have all these options you know you can do the targets individually or various orientations or saturation link but a biparatopic you're both arms of that binding domain are bind to different parts on the same molecule and if it's working right they do so simultaneously so those affinities

53:05

are essentially always aidities like there it is you're the combination of the combined affinity so in that case um it's almost more of a profiling than a tight or discrete kinetic description of a 1:1 binding because it is not a 1:1 binding so sort of the conventional understanding of Ka and KD don't truly apply so I think if you have knowledge of the kinetics of the individual binding components and you inject the biparatopic over the analyte or the analyte over the bipartopic you should see a significant difference in that binding affinity probably a much more stable binding but it may not be a totally discreet affinity although very often a a bipar topic will look like a one to one interaction it is just much higher affinity than the individual component so often times they may behave that way because there isn't um the the interaction can saturate not like if you were doing a an avid analyte

54:08

over a surface it's a little complicated but um yeah it it it is an interesting phenomenon where often times comparing them to their parental controls important to demonstrate that there is the enhancement of that binding affinity and increased stability um but it may not fit an exactly one to one interaction because it is a sort of different entity at that point but unless you have sort of truncated antigens where the binding domains can be accessed discreetly you know like um chimeic antigens or or clipped or or truncated forms that only access one or the other binding sites um you might end up with complex kinetics

54:48 Cheri Salazar: okay thank you um I would like to thank Dr Bedinger and Dr Xu for uh your answers um they were great uh but that's all the time we have for questions um as I mentioned earlier if we didn't get to your question we will follow up with you after the webinar thanks to all of you who attended this webinar uh if you have additional questions you can follow up with us at questions@carterra-bio.com a recording of this webinar will be available on demand on our website tomorrow thank you again and have a great rest of your day bye-bye now