View original Presentation

Nadiezda Fernandez Oropeza:

Okay, good afternoon everyone. Hello, my name is Nadiezda Fernandez Oropeza, and I am the US Field Application Manager at Curiox Biosystems. I thank you for joining me today to talk about the importance of sample prep, and the effects it can have in the quality of your data.

Let me start by giving some background and some motivation that we have to develop, what is today’s Curiox technology? As you all know, in the past few decades, we’ve seen biotherapies come a very long way. We’ve seen great development in immunotherapy, cell gene and therapies, and vaccine in general. And we have also seen tremendous progress in all the analytical tools they rely on, flow cytometry in general, high color content cytometry, mass cytometry, and nowadays, single cell sequencing. But there is a part of this process that still remains a little bit outdated, and that is the sample preparation portion of things, which relies on something very outdated itself, which is centrifugation.

What happens with centrifugation is, it demands many manual intervention steps. What manual intervention steps bring to the table is that they can contribute to what is known as, “The reproducibility crisis.” And here, we have a very nice diagram that we adopted from Actoris, that gives us a good depiction of what happens when we have a singular intervention points, and what they can compound when trying to measure pretty much anything. These are small choices that we make day-to-day, like buying reagents from different vendors, the way we flick our samples, et cetera. And it is true that these things will only account for minute variances in your measurement, but the problem is, when we start compounding them and adding them up, you can see that this can result in a wide range of potential outcomes that will be very different from your true value. When you’re far away from your true value, unfortunately, you might end up putting out there some results that are inaccurate. It can affect your decision making, and it can also have effects, either costly, or making your process a little bit longer than what you would like it to be.

So a good example of these small choices that we make is centrifugation itself. And as all of you know, centrifugation has some hurdles, like we probably will be pelleting some of the debris along with our cells, but it also requires many manual human steps, aspiration, flicking, re-suspension. And each one of these steps brings with an inherent variance on its own. So, at Curiox, we were wondering, “What can we do to automate this part of the process?” And this is when we would to propose the Laminar Wash technology as an elegant solution to bring your sample prep to the current century.

Now, you might be wondering, “How is this so different from conventional centrifugation and conventional washing?” So as you all know, when we rely on centrifuge-based washing, at its core, what we have is turbulent forces, very high forces that are sometimes up to 500 g, that can be really harsh on your samples. On the other hand, you have repetitive steps that you have to carry over and over again, and that can make your whole process very time-consuming and labor-intensive. And at the end, you have also the issue when working with low cell counts. Now, you cannot see the pallet and it’s very difficult to aspirate the supernatant and re-suspend the sample. That could lead to cell loss and inaccurate data in general.

So in contrast, we have developed this technology that relies on very gentle forces, like gravity, laminar flow, and it can effectively remove all the debris from your sample, and it can do it in a very precise manner, automated manner, and a timely manner. Now, I’ll explain to you how the technology words. I would say that, at its core, the technology has two main components. One is the plate, the other one is the washer. Let’s start with the plate. So for the plate, at a first glance, you can see that it has a conventional 96-well format. But unlike conventional 96-well plates, it does not have physical walls separating the wells. Instead, we have a hydrophilic well, in blue, surrounded by a hydrophobic area, they’re surrounded in pink. So these wells, hydrophilic wells, can allocate your sample, your buffers and your reagents. Meanwhile, all the hydrophobic area around them will repel the liquid and enable droplet formation. It’s also this droplet formation that will allow the samples to be separated from one another.

So after you added your reagents, you added your cells, the next step is incubation. So during the incubation time, what will happen is, the reaction will be taking place, but at the same time, you will help your cells, that are heavier than the medium, to descend down to the bottom of the plate, only by means of gravity, while the lighter and more buoyant particles will remain at the top of the droplet. So you also can see here, in this nice depiction, is that we have a formation of a layer of the cells at the bottom of the well, which differs tremendously from the harsh pelleting that we see with conventional centrifugation.

After you finish your incubation period, it is time to wash, and that brings us to the star of the show, which is the washer itself. The washer, it’s very easy to use. All you have to do is, bring your plate, customize your flow rate and the number of times you want to wash, and hit, “Start.” Once you do so, you will see that the machine will bring a fluidic head containing your wash buffer of choice, and with the head, it will bring two nozzles per each single well. One of these nozzles is in charge of dispensing fresh buffer, in this case, it’s going to be the one to your left, and the other set is in charge of aspirating all that waste one, the on your right. This generates, as you can see, a laminar flow, that has the highest speed at the top of the droplet, while the velocity of the fluid at the bottom of the well, where your cells are, is close to zero.

Something that this laminar flow will introduce too is diffusion. And with diffusion, we will bring all the unbound material back up to the top of the droplet, and that can be washed off again. Once you have finished this, you have your cells ready, re-suspended in 25 microliters, and you’re ready for your next step, whatever that may be.

So, so far, I have described very interesting way on how you can automate your washing process. But what happens when you want to take this one step further, once you want to automate your whole workflow, what are your options? So if you go with centrifuge-based methods, you basically have two different options seen here. One is semi-automated, where most labs will have automatic liquid handling systems, but to do the washes, they still rely on manual intervention. They will take the samples out, wash it themselves, and bring it back to the liquid handler. Or you can go with a fully automated system, which is a very convoluted system because it’s costly, sometimes very difficult to integrate. And you can see here listed, I won’t read them for you, but some of the major hurdles, like space, or the aspiration of the supernatant can be really, really troublesome. Once you have a system that is so specialized and so complex, that also leads to having to do very heavy maintenance on daily basis.

So at Curiox, we think it is time to do something different, and we are bringing our solution to automating your workflow, and that is called the Auto1000. You can see it here, and you can also see it in action on our booth if you would like to stop by. So it is a Hamilton NIMBUS robotics system that has our washer already integrated in the deck. This can be connected to up to six different wash buffers, whatever your protocol might require. It has precision and accurate pipetting. And something we are very proud of is the laminar wash auto software that we developed, that our engineers developed. It’s a very intuitive [inaudible 00:09:20], very easy to learn, we have trained many of our customers already. You don’t need any script background to be fluent in this just in a couple of days. And more importantly, you will provide some traceability for your experiments. And of course, it has an optimized deck layout that contains everything that you need already enclosed in the NIMBUS system.

So, I will just summarize this really quickly. I will say that we developed this technology as a turnkey solutions to enhance efficiency in your workflow. And we are very proud to say that, nowadays, we have many customers that are established companies, established institutions here in the US, who are already generating data for us. And most of them have their own flow experts, I should say. And now, they evaluate things like staining indexes, debris removal, cell retention, et cetera. So if you have any questions, please come by our booth, we’ll be able to show you more, share more data with us, and we’ll be happy to answer any questions. And now, for the next part of the session, for the next half, I am honored to invite Dr. Ben Hoffsman from UCLA, and he’ll be describing his experience as an evaluator of the Laminar Wash technology. Thank you.

Dr. Ben Hoffsman:

Is this going to automatically… Let’s see. Get this queued up. Do you want to just pull that up?Great, thank you.

And I want to just thank the organizers at Curiox for inviting me to come and speak here today. Melvin here, you sent me a draft of the abstract and I think you used the word, “Incremental” in there, which I kept. And I think that it really speaks to the evolution of antibody discovery in my experience, which started…

I’ve been at UCLA for two years, but I first was introduced to making hybridomas back in 1991 with this woman named Elizabeth Wainer, you probably haven’t heard of her, but she discovered Tysabri, which is a drug, antibody, therapeutic licensed by Biogen [inaudible 00:11:58] for MS and Crohn’s disease. And I learned the Cesar Milstein classical PEG fusion-based hybridoma generation in her lab. And, of course, did these more classical ELISA-based which we still do today, screening with Liz.

And then, went on to do my PhD at UVA. And I was remembering some of those transitions, it was really frustrating at the time because, when I was working with Liz, we were basic scientists, we couldn’t get a lot of clinical tissue and samples from the clinicians. So when I went to do my PhD, I went and worked on developmental biology, so I had plenty of frog samples. But learned quickly that I wasn’t going to retire, like Liz did, off of making antibodies to frog protein. So then, went on and did some small molecule work, and then ultimately, I was recruited by the Fred Hutchinson Seattle to come back and do antibody discovery work.

So fast-forward up to 2010, 2011, I was a director at the antibody core at Fred Hutch for 10 years, and we ran about 30 to 40 campaigns a year there. And in 2012, somebody showed up at the Hutch from IntelliSite and introduced us to their instrument for doing high throughput flow. And we weren’t doing a lot of flow at the time, but we were doing a lot of ELISAs. And we basically worked together with them and developed a lot of the B technology that people are using now on that instrument and other high throughput flow instruments, for essentially coupling proteins to beads, and that’s shown here.

Let me see if pointer’s working. As these populations, I don’t know if you can see that, but the populations on the left, where you can basically gate out these populations and then look at antibodies binding to those beads, and you can do multiple targets simultaneously. And when you look at the data that’s coming off of here, you see, these are just the MFI signals, you can see that you got this pretty large dynamic range, in some cases it’s in the millions, in the background, tends to hover in a few thousand.

So we went ahead and did these bead-based screens, and we were doing cell-based screens for several years there. And just to highlight some of the highlights or the benefits of the beads over the traditional ELISA there. This is a screen we did at the Hutch, it was for Stan Riddell, who’s a very prominent CAR-T person at the Hutch. And this was for a biomarker. So we started off, the upper right-hand sequence there is the cytoplasmic tail of ROR-1, and the goal was to make antibodies to that tail, which we did. And then, in the yellow box there, you can see we had around 50,000 clones that we screened, 1200 positives, by this high throughput flow on beads. And then, we jumped down to about 100 or so that we ended up actually screening by IHC, that ultimately gave rise to the 64, which was the molecule that was actually licensed out as a biomarker reagent.

So two things I wanted to point out. One is the jump from the 1222 to the 125. And then, the other point, when I was putting a slide together, I was remembering is, when we were doing a lot of the antibody work for people, they would come back to us, we would hand them a list of clones and we’d be like, “Here’s your good ones and then here’s some extra ones.” And they would often come back and say, “Well, I don’t really like the good ones but I like the extra ones.” And after a while, we realized, “Okay, well, it’s because we’re doing these high throughput screens and we’re not washing it all.” And that was really the only way we could get through literally tens of thousands of clones within reasonable timeframe. So then, when you go back and do a more traditional flow experiment and you saturate with your primary, wash that away, saturate with a secondary, wash that away, the order changes around.

And so then, jump forward to 2020, over the course, while I was at the Hutch, I was working with another very prominent physician scientist, Dennis Slamon, who’s at UCLA. And Dennis, after proving ourselves on a number of successful projects with his team, ultimately invited me down to UCLA, and that’s how I managed that transition. But we were designing this, which is our current version 2.0 automation platform for antibody discovery, and we had a centrifuge on it, because we wanted to centrifuge the samples down and ultimately wash them as part of the workflow to reorg and get the true positive signals and reduce the background that we were seeing in the no wash screens.

And fortunately, I actually reached out to a colleague who was, at the time, at Curiox, and introduced me to their instrument. And we downloaded it, and we looked at some results comparing basically beads that were centrifuged, versus beads that which were washed by laminar flow, and we saw things like this. So essentially, the clones are going across the bottom, and then the MFIs are going up. And the green bars here show this sampling of clones with varying degrees of MFIs, with the main one being further on the left. Then the blue ones are the ones at that we did with… Sorry, let me… The green ones were no wash, so non-centrifuged. The next slide has to do with centrifuging. So these were no wash versus laminar wash. And you can see that, what I was mentioning before is that, the order of things really changed around. So now, we’ve got this middle clone 3D2 or 3O2, which is now really showing a better signal-to-noise over the clone that was first identified without washing. So that was the first thing.

Then, the next thing we wanted to do is look at how cells behaved in the same format. So we took a quick look at the gross morphology of the cells, we didn’t see any major changes, the viability seemed similar. Now, this is actually with washing. And if you look then at the MFIs of those, and I ordered this data by the ratio. The blue column there that’s highlighted in green is the order of the clones based on washing by centrifuging. And then, if you could look over on the far right, you can see that the ratios over here change. So a lot of the order, essentially, we were seeing, was different. And we think that this now is really something to do with, what we think is, if you can imagine more physiological conditions as opposed to the harsh treatment that the cells actually go through in the centrifuge.

And I’ve got one other piece of data here, this is just going further down the list on the ranking, and I didn’t highlight there, but you can see that, again, the centrifuging hierarchy is quite a bit different than what the list that you get when you’re washing by laminar flow.

So with that, I’ll wind up, and I think the red light’s going off here, 26 seconds over. But I’ll take some questions, and Nadie, if you want to come up in field as well. Thank you. Yeah?

Speaker 3:

So compared to centrifuge washing, how similar were the washes? Did you have to do more wash cycles to clean up your cells similarly?

Dr. Ben Hoffsman:

Yeah. You might have your own data set. We did, I don’t even remember, Mark might even remember what we set up, I think we were doing two or three washes by laminar flow, and I’m sure we were doing some minimal number by centrifuging, just because it’s so much more involved. So I think they’re very comparable. And you can imagine the time involved in re-suspending pellets, and I didn’t really go into that, but it was pretty obvious from what Nadie was presenting today.

And I should mention, I failed to mention this, but we did, you saw from the workstation, we have integrated the instrument, so we have it as a full automated integration workflow.

Speaker 3:

Nice.

Dr. Ben Hoffsman:

Numbers-wise it’s comp… I think the actual buffers are that the cells are seeing as comparable.

Speaker 3:

Okay.

Dr. Ben Hoffsman:

But the time is obviously a lot faster.

Nadiezda Fernandez Oropeza:

Just to… I’m so sorry.

Speaker 3:

Oh god, no.

Nadiezda Fernandez Oropeza:

I’m sorry. So just to give you an idea, we have suggested number of washes for every single application, but roughly speaking, one of our cycles takes about 22 seconds to be completed. For most immunophenotyping assays, we recommend nine washes, so we’re looking at a whole wash process being completed in around three to five minutes. So it’s something really quickly, and with that, you get good debris removal, and a whole exchange of buffer from the old one to the new one.

Speaker 3:

Okay. And I have one more question. How long does it take the cells to settle on your plate in comparison to centrifugation?

Nadiezda Fernandez Oropeza:

So we use, as I was showing the slides, I don’t know if I mentioned this, but we use any conventional incubation time, like you would do antibodies with your cells, also as the settling time, so you don’t have to add extra time for the settling. Roughly speaking, we need somewhere in between 20 to 30 minutes to have all the cells down at the bottom. So that’s usually the time that most scientists spend incubating in any given process.

Dr. Ben Hoffsman:

I can just add on, the data I showed on the cells, those were actually fixed cells, and if you’ve worked with fixed cells, they become a lot more buoyant and lighter, and we didn’t have any problem with those settling out in washing. There’s a few cells that you’re going to lose, but if you have a reasonable number, the data’s really just fine.

Speaker 4:

Hi, thank you for the presentation. We are a also user of the Curiox and sold on it. One of the stages we’re at right now is, we have done our validation of using this tool and replace of the centrifuge to find some of our target cell top populations for screening, and the impact has been so great that we’re actually trying to understand how else we can use the tool. So for our application, it’s for immunophenotyping tumors, and I’m wondering, in your case, for how you’ve done the screens and using the beads as a way of discovering how this is a valuable tool for you, how are you using it as an innovative tool, if you don’t mind me asking? Because it does seem to imply you can look at a whole new biology.

Dr. Ben Hoffsman:

Yeah. Yeah, there’s been a lot of… It’s interesting. We’ve had a number of ideas come and go. We are using it for, we have actually grown some cells on the plates, and we are using it in that context. So as opposed to doing flow, we’re doing some other assays. But I probably can’t go into all the details around what the assays are, but we have actually done that, gone that far. But there’s clearly, just the simplicity of the design of the instrument is really lends itself to exactly what you’re saying, and I think there’s other things that it’s very amenable to as a resource. But happy to share that, we’re trying to put this together as a manuscript, but I can talk to you about it later.

Did you have any other?

Nadiezda Fernandez Oropeza:

Nope. Thank you everyone.