Scientist-to-Scientist Series: Characterizing antibody libraries using HT-SPR

0:00:00.0 Alyssa Hughes: Hi, I'm Alyssa Hughes, Application Scientist at Carterra, and I'd like to welcome you to another Carterra Scientist to Scientist video. Our guest scientist today is Randi Westh Hansen, PhD. Randi is a Senior Scientist for Antibody Technology at Symphogen and she'll be answering questions today on characterizing antibody libraries using HT-SPR. Welcome, Randi.

0:00:23.0 Randi Westh Hansen: Thank you.

0:00:25.1 AH: Let's get started. So first, thank you so much for giving such a great presentation at our recent symposia. It was really engaging. Could you describe a little bit on what your lab focuses on and the applications that you're running?

0:00:40.2 RH: Yes. So as you just said then, I'm working as a scientist here at Symphogen in our Antibody Technology team. So at Symphogen, we are focusing on developing therapeutic antibodies mainly within oncology and inflammation. So in our team, we are generating the antibody repertoires, and then perform high-throughput screening to characterize the antibody and then down to the final lead. So I think our workflow is shown on the first slide you have.

0:01:16.9 AH: Yeah, let me get that up. All right.

0:01:21.9 RH: So this is basically showing our workflow in our early discovery process. So we get our antibodies from B cell, from immune donors, and then isolating the B cells and then we generate recombinant antibody libraries by sorting the B cells, and then making the antibodies by our Symplex technology. So here you can see that we start around 10,000 recombinant antibodies. And then when we are about down to around 400 antibodies, we start to characterize the antibodies by SPR where we use the Carterra LSA, and has been working with it since late 2019.

0:02:13.5 AH: Oh, wow. Yeah. That's a big jump in your screening process for sure. So what is the average timeline through that process with the LSA? What's the average time you're spending on HT-SPR going from that 400 to one?

0:02:31.4 RH: I think it really depends on what kind of assays that we perform. So on the next slide, there's an overview of the different assays that we do when we characterize the antibodies by SPR.

0:02:46.5 AH: Okay.

0:02:47.9 RH: So we do... In the beginning when we are around the 400 antibodies, we do the affinity ranking and also of course the binding kinetics of the antibodies, and also do a cross-reactivity to all the [0:03:00.1] ____ antigens. And I would say that these basic affinity assays, in principle, we can perform it within a day or two, but if you need to have some optimization, and also the analysis itself, I would say that you need at least a week, maybe two weeks, also because often we have other tasks to do, but the focus time on the LSA, that can be done in a week.

0:03:30.7 AH: Wow, yeah. That's incredible.

0:03:33.5 RH: Yeah.

0:03:35.4 AH: So can you describe a bit about some of the epitope work that you do and mutant mapping data, and things like that and how that impacts your decision-making for antibody selection?

0:03:48.6 RH: Yes. So we both do a epitope binning with this competition assay where you get an idea of how the libraries, how they cover the epitopes of the antigen. And then later on in the process when we have identified the final lead, we have this epitope mapping where we take the antigen and then mutate all the positions of the amino acids. So do a [0:04:15.6] ____ scan and then measure again, the antibody, where they start to lose the affinity for the antigen and thereby characterizing the epitope of the antigen. So I think in the epitope binning there, we will get the broad overview of how our antibody covers the antigen, which we then use in the selection if we want to target a unique epitope and also differentiate for if there's any competitors. And then for the epitope mapping of the final lead, we mainly use that for IP. So that is to have a stronger patent when you can characterize the epitope as well.

0:05:00.1 AH: Okay. Yeah, that's... Yeah, your work is incredible. It's very cool to kind of down in the details of that. You mentioned a little bit more of your cross-reactivity work. Could you kind of dive into that a little bit, and describe how you screen for cross-reactivity on the LSA?

0:05:21.6 RH: Yes. So we do it as a first pass affinity, we call it. So it is when we are here at 400 antibodies. So this example here shown on the slide, I think it's around 200 antibodies. But when we're setting up the affinity assays, we set it up so in the full assay that we first capture the antibodies and then measure the affinity to the human antigen, and then regenerate the assay and capture antibodies again from the same plate, and then measure affinity to, for example, the cyno antigen and then again to the mouse. So in one assay you can get both the affinity and also the cross-reactivity to, for example, these three antigens shown here on this slide. So here we can use it for our selections if we want to have cross-reactivities to, for example, cyno and mouse, and then we can already early on in our projects, we can start to design the in vivo assays later on. And also if we have any talks, we always have talk studies, but also so we can prepare for the talk studies later on.

0:06:34.6 AH: Okay. Yeah, so cool. Okay. So last slide that we've got to cover with you. Can you dive into a little bit more on your high-throughput epitope binning? Specifically if you're losing 10%-20% of your antibodies, or they're becoming inactive during immobilization, how does that affect your data analysis?

0:07:00.7 RH: Of course, then we will not have the epitope information on those antibody. But again, I also think here that it can be project-specific how many antibodies we actually lose because we often do this high-throughput epitope binning, as we call it, quite early in the projects as well. So it's not always that we have deselected the antibodies with loose binding. So this can also be part of it. But I still think that this, that we're losing 10%-20% upon immobilization. Then we will not have the data of these antibody and get this idea of the epitope coverage. But again, we have a lot of other data. So we have both sequence data and functional data and binding data. So still sometimes also from the sequence data and clustering of our sequence data, we will have an idea of where we think our antibody binds. So often we do not use that much time on optimizing the conditions early in our projects. Yeah. [chuckle]

0:08:13.5 AH: Yeah, yeah. And it seems like that really kind of improves the quality of your analysis too kind of narrowing things down. So other questions that I have, can you kind of dive in a little bit... Again, your presentation at our recent symposia, viewers are getting kind of a snapshot of some of the things that you presented on. Could you kind of describe some of the key attributes that you uncover using the LSA to identify therapeutically relevant antibodies?

0:08:47.9 RH: I think that the key contributes from the LSA is that we quite early in our screening processes can get the kinetic binding information of our antibodies. And also this epitopes binning I think that's also have a very high value to our projects and especially... And this also work with bispecific antibodies where you really want to cover and test different epitopes on the antigens, and then see how they can work together. Then here, it's having a high value that you can get some of this information early on in the projects.

0:09:29.5 AH: Yeah, yeah. That's really great to hear. So Randi, I wanna thank you so much for participating in our Scientist to Scientist series. I know you're super busy. We appreciate your time and input in helping us develop more content that can assist others in understanding high-throughput SPR and how it's working in real labs around the world. Again, thank you so much for your time. We really appreciate it.

0:09:55.7 RH: Thank you. You're welcome. [chuckle]

0:09:56.6 AH: Yep.

0:09:56.8 RH: Thank you.

0:09:57.3 AH: All right. Have a good one.