[00:01:06] Aleksandra: Welcome to the digital pathology podcast.Today my guest is Sharla White. She is the vice president of research and development at ClearLight, a company that enables scientists to see more biology in their tissue samples, through tissue clearing and the 3D immunofluorescence. This is a truly fascinating technology and I’m super excited to have her as a guest today.
[00:01:34] Hi, Sharla. It’s great to have you on the podcast. Thanks for joining me. How are you?
[00:01:39] Sharla: I’m doing good. Thank you for having me. This is always a pleasure.
[00:01:43] Aleksandra: I’m glad, I’m glad, and we’re in different time zones, so I’m glad we found the time also to match the time zone. And let’s start with you. Tell me about yourself, about your background and how you joined the company.
[00:01:59] Sharla: Oh, sure. So I have a bachelor’s degree in chemical engineering from Washington University, my love of math and science, it seemed like the perfect mesh, but then along the way, I decided I wanted to get a… Ended up getting my PhD in pharmacognosy. So that’s the study of natural plant products, and pretty much bringing this scientific validation to active components that can be found in natural plants.
[00:02:29] So I pretty much took a hard left into the biology territory instead of chemistry for that, and that was very interesting and engaging, and I was very excited to fold that knowledge in. One of the favorite things I like to tell people is, aspirin really was identified from willow bark. So that’s the foundation that makes it easy for people to understand what pharmacognosy was, because often what you get is, “Is that a real word?”
[00:03:03] So I spent five years doing that. I had a great experience there, and then decided to do a postdoc Stanford University, and was working with a cardiovascular surgeon who was looking at a resistin, or she was looking at ginseng and the natural component, resistin, that could be used for treating restenosis and cardiovascular disease.
[00:03:27] And then that turned into vascular immunology focus, and it was time to move on, so then I joined Genentech and it came a little bit full circle for me, not in a nice, clean way, but I was then doing cancer immunology. And it was ironic because actually, my graduate advisor had two different projects that she was looking on, and one was focused on women’s menopausal symptoms, which is what I was doing, and she had a heavy emphasis on breast cancer, and particularly looking at chemical aspects for that.
[00:04:05] So now I found myself back in the cancer world for that, but bringing in this immunology focus that I had from primarily working with human samples, which was the gold standard when you’re doing research, if you’re doing preclinical, always that, “Okay, we can get some animal samples. It’s great, but what you really want are human samples,” which are just unattainable unless you’re doctors.
[00:04:27] Then I went to the other realm of that, and then I found myself back on the preclinical side of it, which was still interesting, but with a new focus. So I was really excited about that, and then while I was there, I happened to hear about this opportunity with ClearLight, who was taking this technique called CLARITY, that they had been applying to the brain, and all I could think is, “Wow, some of those images,” if you’ve ever seen Carl’s original CLARITY paper from 2013, 2014, it’s… “Wow, this is amazing,” and-
[00:05:02] Aleksandra: We’re going to link them in the show notes, this paper.
[00:05:06] Sharla: Okay. Oh, great, because it’s [crosstalk] .
[00:05:07] Aleksandra: We can do that. So if you have resources, throw the resources in the podcast, and everything that we can link, it’s going to be in the show notes.
[00:05:16] Sharla: Excellent. So we saw this paper, or I saw this paper, it’s amazing, and then there’s videos where they play through it. And the whole premise behind ClearLight was that we want to take this technology and apply it to cancer. All I could think was, “Yeah, this would be amazing,” because there’s so much going on that we think we know, but then we don’t know. And cancer seems to be this ever-evolving living thing, obviously, because they’re cells, but developing escape mechanisms, and just throwing up flares to distract. So I was like, “This would be great.” So I ended up taking this opportunity at ClearLight, and I joined as a scientist, and spent the first several years trying to understand and adapt CLARITY for tissues that weren’t the brain.
[00:06:11] Aleksandra: So let’s take a step back and let’s explain the technology first.
[00:06:16] Sharla: Okay. So CLARITY is one of several different tissue clearing techniques that are available to researchers, the idea being that in order to… We want to maintain the integrity of biological tissues. So, we are in 3D. Cells are not flat, 2D surfaces, they’re in 3D. But often when we’re trying to image these intact tissues on a microscope, we’re not able to see it. It’s very hazy. Maybe you don’t even get to penetrate it.
[00:06:51] So the thought behind CLARITY and tissue clearing is that in order to image deeper, what you need to do is remove the light-scattering lipids. So you’re basically aligning… You remove the light-scattering lipids and then you align the refractive indexes of collagen and bone and all of the different components that might be present in any given tissue.
[00:07:14] And once you do that, you’re then able to image deeper into the tissue, so beyond the surface of like 100 microns, and then you can actually see… And then the other part of this is, you have your hydrogel scaffold. So we take a fixed tissue, we embed it with this hydrogel monomer, and then we polymerize it. That establishes these cross links, and these cross links create a support scaffold for the tissue or the organism that you’re looking at.
[00:07:44] And then we go on with the detergent and we remove the light-scattering lipids. So that aspect is what allows us to image deeper into the tissue. But the key part is that you want to be able to maintain the structure. So we remove the light-scattering lipids, and now we’ve got this hybrid tissue scaffold instead.
[00:08:02] So we’ve blocked the proteins and the RNA and the DNA into place, and then our next step is to then do immunostaining. So this 3D IHC technique that we’ve developed, so basically it’s making sure that you’re using an antibody that is compatible with the process, because not all antibodies are created equal, but also that you’re using the right concentration for even, uniform staining throughout the entire tissue so it’s not going to congregate on the surface. You haven’t used too little, where it peters out by the time you get to the middle.
[00:08:36] Aleksandra: How do you get them to the middle?
[00:08:40] Sharla: We’ve done a lot of optimization on the process, that’s one. But the other part that was one of the key things that we determined is, you have to make sure you’re using the right antibodies. So if you’re doing like western blot, or even just standard FFPE, we’re usually using [1:10]00 dilutions, maybe [1:20]00, super… We’re able to stretch that vial far.
[00:09:01] But because we have so much more tissue, you need so much more antibody. So what we found… And some antibodies, you don’t need to go as high, but if you don’t have enough, then you’re going to get so far and then it’s just going to be gone. If you don’t clear it, and it looks sharp, and then all of a sudden it’s like you’re… It’s when you’re reading a book and you look up real quick and it’s a little blurry, that’s what it looks like, if you’re imaging too deep and it’s not really clear, all of a sudden everything becomes hazy, it’s not in focus.
[00:09:33] So we have spent several years, one, understanding that aspect, but two, trying to find ways to improve that staining, improve the imaging aspect of it, because that’s its own kind of nuanced approach for it, trying to find the best microscopes that will allow for that in what we’re doing.
[00:10:00] Aleksandra: Yeah, that’s another question, how do you image that? I mean, it’s basically 3D IHC, right?
[00:10:07] Sharla: Yes.
[00:10:08] Aleksandra: So anybody who knows how complex one-dimension IHC is, is going to have the same question, like how do you get it inside? How do you image it? And like you said, it’s a lot more antibody because we have a 3D structure, and you said you start with brain, which makes perfect sense to me now, because you said you’re removing lipids.
[00:10:27] Sharla: Right.
[00:10:27] Aleksandra: This is the tissue that is like 90%… I don’t know how many percent, but a lot of percent lipids.
[00:10:34] Sharla: Exactly.
[00:10:34] Aleksandra: So when you remove that, you basically have something that is easy to penetrate.
[00:10:38] Sharla: Yes.
[00:10:39] Aleksandra: How do you do it for other organs?
[00:10:42] Sharla: So, with a lot of failure. We spent a lot of time trying to make the other organs look like the brain, which in hindsight seems silly, but we really just had to establish a foundation in understanding what exactly the technique is doing. So, with a kidney, there’s going to be a little more structure, there’s vascularization, right? Which aren’t made of… So what happens is, you remove some, and you’re able to start to see the vessels, and that actual vasculature that’s occurring, but that’s never actually going to clear itself out.
[00:11:22] And then it gets even more complex when you start talking about these heme-rich tissues. So you’ve got a heart or a liver or a spleen. So now you’ve got to work with the autofluorescence that they all tend to have, and then what do you do about the heme that’s there? Do you want to decolorize it? How do you treat that and maintain the structure? And then [crosstalk] –
[00:11:46] Aleksandra: But then your chemical engineering background comes into play very nicely here, because this is very technical.
[00:11:52] Sharla: It was not intentional. Before I took this job, I considered myself a jack-of-all-trades molecular biologist by the time I had finished my graduate degree, but I thought that primarily because I wasn’t really doing as much chemistry that was going on. But here we’ve got the chemistry that’s involved with the hydrogel, and you need to be able to improve.
[00:12:22] So we pretty much focus on the passive process right now. There’s been a lot of emphasis on doing the electrophoretic clearing, and that’s where you pass the current through, and the goal is to speed the process up. And we’ve done some preliminary work in that aspect too, but the back end of it is, most people don’t have any kind of specialized equipment to be able to do it, but they want to be able to look at tissue clearing. So, with passive clearing, you don’t need that special equipment. You just need to know how to embed your tissue, and how to remove the lipids, and the optimization for your antibodies and all that, and then just have the right microscope in place.
[00:13:06] So, how do you improve that process? And that’s where we start making little tweaks and changes to the chemistry of all of the different reagents that are at play, and so I’ve ended up doing a fair amount of research on that. So now the chemistry comes back into play and it’s less straight chemical engineering and more biomedical, biochemical, which is great because now it all is like, “Oh, all of these skills that I picked up along the way are really paying off right now in this moment as we try to solve these problems.”
[00:13:46] Aleksandra: Good. So you started as a scientist and now you’re the president of research and development at Cleveland?
[00:13:53] Sharla: Vice president. Yes.
[00:13:54] Aleksandra: Vice president. Vice president of research and development. Yeah, so that’s also-
[00:13:59] Sharla: How did that happen? I-
[00:14:00] Aleksandra: And then I was going to ask you about the… We’re not done with the technology. I mean, still, we have to talk about the company, because that’s interesting, but I’m a pathologist, so I’m definitely interested in the microscopes you’re using and the imaging. Let’s do that.
[00:14:15] Sharla: Yes. Okay, so-
[00:14:16] Aleksandra: And then we’ll go back through what you actually do in the company as the vice president of research and development.
[00:14:22] Sharla: Oddly enough, all of these things have worked their way in to even get to this particular point, but okay, you’ve got your sample, you’ve cleared it. Now it’s some level of thickness, and there are two main imaging platforms for doing this. You’ve got confocal, which pretty much is the workhorse in any lab, if you’re lucky or you have access to a core center, and then you have light sheet, which is… So they’re two different things. With confocal, you’ve got a pinpoint, so you’re illuminating the entire tissue in this one particular point, and it lights up everything. The resolution is great. It’s nice and crisp, if everything goes well. Fantastic.
[00:15:11] With light sheet, what you’re doing is you’re passing a laser sheet through one plane of the tissue, and by doing so, it allows for faster imaging acquisition and less photobleaching, which is the downside to doing confocal, but you also… Most of the objectives on light sheets have a longer working distance, which basically is how far away from the top of the objective glass can you image into the tissue?
[00:15:40] Trade-offs are usually… So now I start learning about numerical aperture, which is how sharp your image is going to be. All of these things impact the resolution. Everybody wants to be able to have the confocal resolution without having to dedicate the confocal time and not the photobleaching.
[00:15:59] So we, because our original impetus was to focus on cancer, we’ve always wanted to focus on the subcellular aspect of it, so 20x. We used to think higher, but just for what we’re doing, it’s just too much information to really go that much higher for large cancer tissue and be thick.
[00:16:20] But we spent years looking at different light sheets before we were able to identify one, like, “This does what we need. It seems to check all the boxes.” And then we came to a new crux, which is, tissues that are clear enough for a confocal may not necessarily be clear enough for a light sheet, and that comes back to where the laser’s going.
[00:16:43] So, by shooting that laser through that one field of view, and you’re lighting everything up, it’s great. That’s how you get the great image, but it’s a different bar than if you’re trying to get an entire sheet. So the entire tissue must be completely clear in order for the sheet to stay straight so you can capture that image. So we have done a lot of work in understanding those nuances.
[00:17:09] Aleksandra: So a sheet would be a counterpart of a tissue section, right?
[00:17:13] Sharla: Yeah.
[00:17:14] Aleksandra: Okay.
[00:17:14] Sharla: But the trade-off is, any microscope is still only doing one field of view at a time.
[00:17:18] Aleksandra: Mm-hmm [affirmative].
[00:17:19] Sharla: So that’s going to depend on your magnification that you’re using. So obviously, the higher your magnification, the smaller your field of view is, and then… But most of the-
[00:17:27] Aleksandra: Okay. So it’s going to be like a field of view in a microscope, right?
[00:17:32] Sharla: So that’s always the interesting thing, when you’re like, “So, hey, look, I know you’re… Just so we’re clear, all of the microscopes, despite their differences, they all acquire the exact same way,” which is, you have to acquire in a rectangle or a square, and they all end up that way.
[00:17:50] So that has been a key part of it, so finding that balance of what’s clear, what’s clear enough for a confocal, what’s clear enough for light sheet. Ideally it’s clear enough for both, so you can do whatever platform is going to work for you. And then making sure that you have the right working distance.
[00:18:08] So we spend a lot of time making sure that we lay this foundation of, “Okay, you want to see a whole sample. That’s great. We’re going to need to image it,” say, “10x, because your sample is so large,” and that trade-off is, you have a lower numerical aperture for that. Or, “Great…” We have a standard workhorse for our confocal. It’s like 25x with a .95 NA, I think that sounds right.
[00:18:36] And then we have a light sheet where the numerical aperture is 1.0, which is great, but these are really, really specific things. I’ve talked to somebody one that said that they had a specialized 10x objective with a really long working distance, and I think their numerical aperture was okay too, but [inaudible] trade-offs, you got to do what works for you.
[00:18:59] Aleksandra: Yeah. So it’s everything, it’s optical physics, it’s chemistry, it’s biology, immunology, basically-
[00:19:05] Sharla: It’s [crosstalk] .
[00:19:06] Aleksandra: … combination of everything.
[00:19:09] Sharla: Which really makes it exciting.
[00:19:11] Aleksandra: Yeah, but you have to be able to look at it from all the angles, or have a team that you have team members that can tackle all the challenges of this. But basically if you’re the vice president of R&D, then I guess you have to have at least an overview over everything that’s going on there.
[00:19:33] Sharla: I think it’s safe to say I pretty much have my fingers in almost all of the aspects of that. And if I’m not doing that, probably because it’s well beyond my skillset. Then my job is to try to convey what it is we’re after that we need to be produced on the other end of it. It definitely wasn’t in the job description when I joined that these things would be happening.
[00:19:58] Aleksandra: So, last thing about the technology. So we said it’s immunohistochemistry or immunofluorescence. You do fluorescence?
[00:20:07] Sharla: It’s fluorescence, yes.
[00:20:08] Aleksandra: Okay, so immunofluorescence, and it’s 3D, so you image it multiple times, right?
[00:20:15] Sharla: Yes.
[00:20:16] Aleksandra: So you have the information from all your markers at multiple planes, and how do you make it 3D? What’s imaging or putting it all together?
[00:20:28] Sharla: Yeah. The trick to doing this in 3D is, we’re going to set it for Z-stack, and on the confocal, for the setup that we have, we’re able to adjust the lasers. If you have the… Depending on the series of antibodies you have, maybe you can also adjust the gains as well, depending on if it’s a non-sequential or a sequential acquisition. But one of the things that aids when you’re going deeper, because it’s going through more tissue, is the ability to increase the laser power as you go along, but not… It’s that balance of the increasing the laser power, but not oversaturating it so you have an area that’s blown out.
[00:21:11] So you do that for your one field of view, but now you have to then apply this to your entire section. So one of the other key things that most people don’t realize is that while you may have signal throughout the tissue, at any given point in time, you may have extremely high expression on, say, the far right side, but at that exact same plane on the left side, there’s signal, but it’s just not bright enough. In order to see the signal on the left, you’d have to increase the laser power, but you would then blow out the right.
[00:21:48] So finding that balance, we have definitely… It’s great when you have a marker that is uniformly expressed throughout, but if it shows up in high concentration, then it’s just going to be bright. So a lot of that is adjusting that, if you have a camera where you’re doing-
[00:22:05] Aleksandra: So like all immune cell markers, where you have clusters of immune cells, that’s going to be the problem?
[00:22:10] Sharla: Yeah. DAPI presents that same thing, when you hit like fat cells, or if you hit a pocket of cells right there too, how do you adjust for that? And a lot of that depends on the specific question that the customer has. Like if they’re like, “Look, I don’t care about this outside part being blown out. I want to see what’s in the center, make that happen,” then we’ll make it happen. We’re like, “We’ll just cut that out so it’s not distracting for you.”
[00:22:36] But we work on that, and if you have a camera where you’re working off the exposure, which is how our light sheet works, then it’s finding the optimal exposure time in conjunction with the right laser power. So we’ve got two different approaches that have to occur depending on the microscope that’s being used, and often how we approach it and how we tackle it is guided by what is your actual research question of interest. So we really try to get to the heart of that when we embark on these projects.
[00:23:15] Aleksandra: Yeah, so that’s also my next question. Who is using it? It seems to me as a super powerful, but also super complex technology.
[00:23:24] Sharla: Yeah.
[00:23:24] Aleksandra: So who are your customers at the moment? Who are you doing this for?
[00:23:29] Sharla: So we’ve got a-
[00:23:30] Aleksandra: And what are their research questions, as well?
[00:23:34] Sharla: In the general sense of that, we definitely have a mix of both industrial and academic collaborators and also customers. We focus on the preclinical RUO aspect of this, and it runs the gamut of what we’re doing. So, for instance, we have one collaborator who, their thought was, “Hey, we like what CLARITY is. We don’t necessarily want to do the immunostaining part of it, but we want our tissue to be cleared so that we can see what’s happening on it under a bright field.” So we’re like, “Okay, we think we can do that,” and so that’s a back-and-forth-
[00:24:19] Aleksandra: So do you stain it with something, then, for bright-field?
[00:24:23] Sharla: Well, we do on our side. Yeah, so for us, we’re like, “Hey, thanks for the tissue.” We can clear it. We can stain it. We can try out new… It’s great for antibody feasibility for markers that will be a positive control for that tissue. So that helps us, especially if it’s a new tissue that we haven’t embarked on.
[00:24:42] For them, we clear tissues and then we send it back, and they work on their side so that they can see if they can track what’s happening now that… You kind of give yourself a biological support that is somewhat transparent, or at least it’s transparent in regards to like a bright-field microscope.
[00:25:00] Aleksandra: Okay, understand, so you basically take care of your part of the technology, and then they take care of trial and error on their side?
[00:25:10] Sharla: Yeah. Then we’ve had other customers where they’re, “We want to see what’s going on in the bone, but,” obviously, “the calcification is causing us problems.” So we have done that. We have a couple of different collaborators that are interested in the eye, and the eye, we’ve been working very hard on making this work for the eye in an intact fashion, because for most of the research that’s out there, if they’re interested in the retina, then they end up cutting it off and then they make a flower petal shape to make it flat because it’s around like a cup. And the majority of the eye is apparently water. So cutting it and sectioning it can be a problem. And then if people are interested in like the lens of the cornea, then they just cut that part off and focus on that.
[00:26:04] So we have spent the better part of at least a year, I think, really interrogating what’s the best way to make this work to really address the question. And it was interesting to us because I don’t know what’s in the air, but sometimes we’ll get a customer that’s like, “Hey, I want to work on the eye,” and then all of a sudden, like three or four other people are also interested in the eye. Or, “Hey, I want to look at the organoid,” and it just seems like a whole cluster or wave of people are like, “We want to do organoid.”
[00:26:38] Aleksandra: It’s the universe validating the research question.
[00:26:42] Sharla: It’s so interesting. And then of course, we get cancer. Cancer runs the gamut. Everybody’s interested in their own particular subset of it, but cancer is really beautiful to image. It really does look like art. We definitely have had moments when we’re imaging some of these tissues where you have your wow moment, like, “Wow, that’s great,” and, “Oh, you can see the vascularization,” and, “Look at how those T-cells are clustering around… What is that? I’m not quite sure.”
[00:27:12] Aleksandra: Yeah, for immune oncology, I mean, the, so to say flat immunofluorescence is becoming the standard for any immune oncology research. But obviously there is so much more information beyond just one section. So any 3D technique is going to be the logical next step. And if you can have it cleared and continues, that’s-
[00:27:39] Sharla: I think it’s been so impactful. And then we’ve had a couple customers that were developing their own antibody, and they wanted to see how that performs with CLARITY. And then we’ve had a couple people that are working on their own medical devices and that type of aspect, which is always interesting because if there’s a device in place, the scientists were cagey, right? Until it’s published, we’re very cagey on the details.
[00:28:10] But you want to be like, “Look, so I really want to make this work, but we’re going to need some more information to be able to know if this is feasible or not. I’m not asking you what it’s made of, like specific… I don’t need the chemical breakdown, but do they tend to absorb light? Because that seems like that’s going to be impactful of what we’re trying to do.”
[00:28:29] Because again, we have these… So melanoma and the pigmentation, it’s great. You’ll image, and then you hit your wall, and then you’re just like, “Okay, so there’s, there’s blackness.” And if that’s the case and we know it ahead of time, then basically we’ll do pre-processing to alleviate that. So that was one of the key things with doing eyes is, that you have to… If you don’t have albino eyes, how do you work around that pigmentation that’s going to be apparent?
[00:28:55] Aleksandra: Yeah.
[00:28:56] Sharla: So it’s really been interesting and exciting.
[00:28:59] Aleksandra: I mean, it’s so logical, when you say it. I would never think of it on my own.
[00:29:04] Sharla: We never think of this on our own. It’s always, “So, we tried this and it didn’t quite seem to work. Do you have any thoughts?” Like for the eye, we… When we’re hitting new targets and new antibodies, often you’re like, “Hey, so what should we expect to find?” And that’s when [inaudible] you’re going through this, you’re like, “Wow, people doing eye, you’re really limited in what you can see.”
[00:29:30] So then it’s, “What do you expect?” type situation, and sometimes you just don’t, and then you start all these other caveats, like, “Oh, here’s an image, but they cut off the sclera and poked a couple holes, and… Okay.” Cancer has always been a little bit more straightforward, possibly because of the background that I have, where you’re just like, “Okay, so if we know there’s a lot of stroma, then we know what to look out for there, and how we can counteract that.”
[00:29:56] But then you start getting into things like bone, and the calcification and, “Okay, did you perfuse? You didn’t? Okay, I understand you can’t perfuse, but just a heads up. You’re going to see a lot of red blood cells. They like to glow everywhere.” So part of what we do on the back end is-
[00:30:14] Aleksandra: More or less the picture, usually it’s on your end to figure out the roadblocks, and you don’t really get feedback on what can cause a problem from your customers because they know this technology is powerful, but there is, so to say… How do I put it? A lot to know about the technology to be able to predict things.
[00:30:44] Sharla: Right.
[00:30:44] Aleksandra: So basically they come to you and they ask you, “Oh, is it going to be feasible?” And then you do back and forth with them.
[00:30:52] Sharla: Yeah, so one of the first things is, we try to get a handle on what have been the issues that they’ve had. Well, okay, so some are, “Look, it’s the eye, we can’t section it, but what we want to see is not on the surface.” So then you can look at it and be like, “Okay.” There’s also a back-and-forth afterwards, like, “This is what we got. Does this align with what you expect?” “No, we want to see this, this and this.” “Okay, let’s fix it. We’ll go back and make modifications.”
[00:31:19] Sometimes it’s, they already know, “Hey, so this is the antibody.” I think we did one and we were like, “It just seems like there’s a lot of background signal, or it’s nonspecific,” and sometimes it just validates what they were already seeing, and they’re like, “Okay, okay, yeah, so we noticed that ourselves and we wanted to follow up and take this different route and re-approach it.” But more often than not, what we get is, “We don’t know what we’re going to find, but we’re interested in seeing what you come up with.”
[00:31:52] Aleksandra: Okay. So it’s not just a method that you apply to what you’re already doing, but rather a lot of scientific study design when working with the client?
[00:32:02] Sharla: Yes. Yes.
[00:32:04] Aleksandra: So what are your top three tissues? I assume brain is the number one?
[00:32:09] Sharla: Brain actually is number two. Number one is going to be-
[00:32:12] Aleksandra: Oh, really?
[00:32:12] Sharla: … any type of cancerous tissue.
[00:32:14] Aleksandra: Okay.
[00:32:14] Sharla: Cancer is one, brain is two. And then third is a close second now, it wasn’t before, but now organoids are our number three.
[00:32:24] Aleksandra: Okay. So cancer, it doesn’t matter which organ the cancer originates from, or do you have like top three there as well, or it doesn’t matter?
[00:32:32] Sharla: We’ve actually gotten the gamut of those. So we’ve done breast, colon, kidney, liver, and did I say lung? And lung.
[00:32:47] And then some of those, not so bad. We internally were initially focusing on breast and lung cancer for a starting point. We’ve done some pancreatic. PDAC is fun, just because of the stroma… It’s a combination of issues when you talk about PDAC, right? So you have the actual, real-life issue of cells themselves being in able to get into it. But from our perspective, it’s also continuing to optimize the reagents that will also allow you to get past that stroma and all this extracellular matrix that exists, for you to really get an idea of what’s going on in that place too.
[00:33:32] Melanoma was its own special challenge, and that is, obviously bleaching is the go-to for anybody that’s doing that. But we had to do several studies to figure out what the optimal conditions for bleaching are, because you can either not do enough or you can do way too much, and you don’t want to leave any kind of biological information on the table because you didn’t source out the optimal settings to begin with. So even with some of these, when we’re doing like our antibody feasibility, tonsil is a great go-to, or spleen, right?
[00:34:12] Aleksandra: Mm-hmm [affirmative].
[00:34:13] Sharla: Especially for the immune markers. But the biggest challenge that we have is less so the tissue sometimes, and more so the markers that people come with, because everybody’s got like a couple of favorites, and then they always come in with one, and it’s just like, “I’ve never heard of that before.” Not that I think I should know all of them, but some of them really just feel like left-field ones.
[00:34:39] When it comes to the brain, what’s become very apparent is that it doesn’t matter how many biomarkers that we have, somebody wants one that we have not done before. Maybe it measures the same things. Maybe it doesn’t. But man, bless the people that are doing neuroscience, because-
[00:34:58] Aleksandra: A lot of respect to them.
[00:35:00] Sharla: … you look at a list and you’re like, “I don’t understand,” like, “Both of those weren’t the same thing? Oh, they’re the same, but different? Oh, okay. Let’s find your marker. It’ll be great.”
[00:35:09] Aleksandra: Yeah. So you say cancer is number one. It doesn’t depend really on the tissue. Is it because of some properties of cancerous tissue, like chemical properties, or is it because you did so much work on this, and that’s why you have this already figured out?
[00:35:25] Sharla: Well, I think it’s just the general direction of immuno-oncology, and the focus of that field to really gain an understanding of what’s going on. So, when I was at Genentech, we were doing [inaudible] mouse studies, right? You’re running them, and flow is your go-to for doing that type of stuff, mainly because you can stain like 12, I think you can get up to 30, maybe it’s 40 now, markers, and you can identify all these subpopulations.
[00:35:55] But what happens sometimes is that you run these studies, and from study to study, even if you’re using the same model, the results can be highly variable, and how do you explain that? Sometimes these subpopulations are really, really tiny, and you’ve identified that 1% out of the other cells, but the process to getting to flow involves you doing a lot of tumor digestion. So you’re losing a lot of cells. So the question from that standpoint, especially if you’re trying to develop drugs, is, your drug working? Is it going where it’s supposed to be going? That type of thing.
[00:36:34] So we did a study with Pionyr. They’ve given us our permission to say their name. But one of those questions that goes across the board for people that are doing cancer research is, what’s the immune response that’s occurring? Now, is it that you want to know what the B-cells are doing, or do you want to know if you’re getting T-cells, be it mature T-cells or cytotoxic T-cells? Specifically, what’s happening with those, and where are they in location to your tumor?
[00:37:06] So studies like that then start to illuminate a little bit more, where if you have… Let’s say your isotype, or your control group, is going to have this heterogeneous mix. That’s fine. We expect that. But then you start doing your treatment, and so we did one of these studies and there weren’t that many CD8s. From a visual perspective, they were sitting on the surface. And when we did the count for the different groups, there was no significant difference in the count for T-cells for them.
[00:37:41] But then when you visually looked at it, you did a treatment, and we see that the T-cells actually started moving into the tumor. But maybe they were only moving 50 microns into the tumor. But then you do another treatment, or a combination treatment, and now they’re fully infiltrating the tumor. So even though the numbers themselves would specifically say there’s no difference in what you’re doing, the combination with the spatial orientation tells a totally different story, which is, “Yes, your drug or your combination treatment is effective, and it’s impactful, and it’s doing this. Your tumor is no longer cold. It’s hot,” and you can’t…
[00:38:24] Often when you’re doing flow, you’re also running like IHC and you’re trying to extrapolate, but you have like your tumor and you’ve quartered it or you’ve cut it in half, and you’re trying to stretch it as far as you can, maybe throw a little FISH on there and see if that’s going to work. And you hope that by putting together all of these applications on the back end of it, that now you have a story to tell. But most of the time, the story isn’t obvious.
[00:38:49] So for us, our thought is, “Let’s help you tell your story in a way where you don’t have to do as much.” It doesn’t mean you don’t do the other things, but all of these techniques have different values that they can add to it. Next-gen sequencing is great, but what might look good for next-gen sequencing from a protein standpoint with IHC doesn’t correlate. So, hey, that’s great, that’s there, but in reality, you might possibly be looking for a needle on a haystack, which, some people want to look for needles in a haystack, and just, hopefully, we ask, don’t send us on a whole organ search to find it.
[00:39:33] Aleksandra: Yeah, definitely. So let’s go back to you. You started as a scientist and now you’re vice president of R&D. Tell me how that happened, and how long did it-
[00:39:46] Sharla: So I’ve been at ClearLight for… Okay, so I’ve been at ClearLight for over five-and-a-half years now. I came in as a scientist, lots of groundwork, legwork, failures, successes. “How do we want to multiplex this? What’s the best way to image it?” And I did that for about two, two-and-a-half years. All the dates run the same for me right now. And then I was promoted to associate director.
[00:40:16] So I work at a startup, so I’ve always had my hands in a lot of different baskets for what we were doing. So, in that time, we started doing some work on the automated clearing platform, an automated staining platform. And then what became apparent while we were doing this is, when we started imaging, there are great softwares out there for doing your 3D renderings of these Z-stacks, but then, how do you get the quantitation on it?
[00:40:45] So I’ve worked with several other scientists throughout my tenure there, and we just couldn’t find anything that did what we needed it to do. So what ended up happening around the time that I was an associate director is, we started to engage that we needed to make our own software. This definitely wasn’t our intent, but in order to get to where we wanted to go, it seemed like it was going to be the imperative step that had to be taken, and oddly enough-
[00:41:20] Aleksandra: So then you add the software development to your expertise?
[00:41:23] Sharla: Yes. I don’t know anything about code, but I know enough general terms that get spit out and what we might want to measure to at least say if it’s relevant or not. But we really had a tough time finding people that were up to the challenge of trying to do this analysis in 3D, because it is [crosstalk] –
[00:41:42] Aleksandra: No, I come from the flat image analysis world, and that’s already… Well, challenge, in a way that when you take people who are outside, who even are trained in computer vision, and then are supposed to translate the computer vision stuff from natural, normal images into biology images, it’s another world.
[00:42:05] Sharla: It really is-
[00:42:06] Aleksandra: Then take it to 3D, and it’s like…
[00:42:09] Sharla: But we eventually got lucky. So then we started developing the software, and now you’re into the question of like, what do we need? So it’s an interesting conversation to have with these engineers when you’re like, “Hey, no, I mean, no, it’s not black and white. There’s gradients for this.” You start talking about like PD-L1, and ER, and you’re like, “No, no, no, they’re not all equally the same,” but these other samples are, so like, “The membrane is, or it isn’t.”
[00:42:37] So trying to bridge that gap where, this is what we need, and this is how you typically approach it, so how do we create a software that works from a biological quantitative standpoint, but also is informative and can learn with these AI and deep learning techniques? And that has been…
[00:43:00] Aleksandra: Oh, my goodness, I’m listening to you and I’m like, respect. There is so much to it. And at the end, you make it work? Respect.
[00:43:10] Sharla: Our biggest win, was how do you get the segmentation for it, right? So our software, we’re still working on getting the alpha, but a lot of the software that’s out there, usually everything’s relegated to like a spot or circle, which, it works, right? But that’s not the actual shape of the cells sometimes, especially when you start getting into-
[00:43:29] Aleksandra: Especially if you’re doing 3D, right?
[00:43:31] Sharla: Exactly.
[00:43:32] Aleksandra: You may have… Like nucleus would be round on a 2D cross-section-
[00:43:38] Sharla: Right [crosstalk] right.
[00:43:39] Aleksandra: Yeah, but then… And lymphocytes, they’re not round little [crosstalk] balls.
[00:43:46] Sharla: Yep, actually-
[00:43:46] Aleksandra: They just look like that at cross-sections. They have like tentacles and stuff.
[00:43:50] Sharla: Yes, [crosstalk] –
[00:43:52] Aleksandra: That’s what they need to do, they go through vessels, so they have to like squeeze themselves.
[00:43:56] Sharla: Oh, yeah, so we spend a lot of time actually educating, to some degree, where there are things that we know, and we know them to be true because they consistently hold up against either the same tissue that’s been provided from multiple sources, or because it’s consistent across all samples that we’re doing it for.
[00:44:19] And then you’re just like, “Look, I know you think that this is what membrane staining is, but you just picked the right level at that…” Like, “How many slides did you go through to find that set of staining? But you go above it and you got the top of it. So it’s not really…” Like, “You think it’s trash, but it’s really not, it’s the beginning of another cell. And then, look, we’re going to go to the next slice, and there it is.”
[00:44:37] Aleksandra: So you have a way to visualize it and show… Okay.
[00:44:42] Sharla: Yeah, so-
[00:44:43] Aleksandra: And now it’s you say it’s a proprietary software, it’s your own software?
[00:44:46] Sharla: The quantitation is going to be proprietary. We license Imaris Bitplane, which is our 3D visualization tool, and so-
[00:44:53] Aleksandra: Do you have video of that?
[00:44:56] Sharla: Do I have a video? [crosstalk] sure.
[00:44:57] Aleksandra: I want to link it in the show notes. If you have any kind of visualization of…
[00:45:03] Sharla: I do.
[00:45:03] Aleksandra: … a 3D image that goes across all the planes, to show the-
[00:45:07] Sharla: Actually, you know what? Yeah, let me… I know we have some on YouTube, and I’m going to find you a good one for that. But yeah, so a lot of what we do, it’s “Hey…” So one of the things that we’ll try is like, “Here’s the 3D volume rendering.” We’re turning the channels off and turning them on, and then we’ll do a slice play-through. So basically, if you were going to be looking at it under the microscope, or if you were going to be looking at it on a slide, this is what it looks like.
[00:45:35] Now, we’re providing a video of that, and you can zoom in certain areas, but often it’s amazing how many people forget that when you do the tile scan of a whole organ, we’re showing you the tile scan of the whole organ, not zooming into one field of view. Because then, if you think about it, if it’s, say, a whole lung, and you want to see an individual cell, we’ve probably zoomed in 400%. So if you’ve got a nice box that says, “Here’s your navigation,” there’s like a spot here that says, “This is where we’re looking,” because we’ve captured so much.
[00:46:12] Aleksandra: Okay.
[00:46:12] Sharla: So, often we remind people-
[00:46:15] Aleksandra: Those files.
[00:46:17] Sharla: Oh, they’re huge. So, let’s see, if you do a whole brain at 5x with a step of 25, that’s going to be about 130 to 150 gigabytes.
[00:46:31] Aleksandra: Okay.
[00:46:32] Sharla: If you do a 25x, let’s say a rather large tile scan, and you’re doing a step size of, say, five, we’re going to be looking at possibly 200 to 300 gigabytes.
[00:46:45] Aleksandra: Oh, my goodness.
[00:46:46] Sharla: And then if you’re on a light sheet, now we’re… The low end is like 400 or 500 gigabytes, and then upward ends is going to be in terabytes.
[00:46:56] Aleksandra: Okay, so we are talking totally different [inaudible] … Pathologists pride themselves that whole slide images are so big, so much bigger than radiology images. Okay, we are talking whole different level.
[00:47:09] Sharla: [crosstalk] yeah. So even if we have these large files and then we make videos, I’ve definitely… If I make enough videos, one video of your sample, just real simplistic, not zooming in, and doing maybe like 2,000 frames or so, is going to be a gigabyte, of a video where it’s been somewhat downgraded. And when we get people… Because occasionally you’ll get some that ask for the raw data, and you’re like, “Hey, that’s great. Can you handle a terabyte of data?”
[00:47:40] Aleksandra: [crosstalk] it on hard drives, or what’s the way of…
[00:47:44] Sharla: If it’s coming off of the light sheet, we don’t even bother to send a link. We’ll do a hard drive.
[00:47:49] Aleksandra: Okay.
[00:47:50] Sharla: For the confocal that are on the smaller side of them, then we’ll send a cloud link that’ll be live for hopefully long enough for them to download it on their side. So that’s the other part that we can’t really account for. It’s just like, “Hey, so we’re going to send you… Here’s 200 gigs, make sure you have enough space wherever you’re going to download it. Make sure that you’re…” We’ve had people where their internet connection… They’ve had issues with the download because their internet [crosstalk] –
[00:48:17] Aleksandra: Download for a week?
[00:48:18] Sharla: … time out, or… Yeah. So I think we’ll probably change that, possibly, going forward. We do an immediate thing, like, “Hey, here go the videos. Here’s your summary report to go with it.”
[00:48:29] Aleksandra: But can your customers even deal with the raw data?
[00:48:35] Sharla: So, we have one customer that can.
[00:48:36] Aleksandra: So, on one, and from the technology standpoint, on the other end, from the scientific information extraction standpoint, how does that work? Or how do you work together to bridge the information that you get from ClearLight to, how do you extract the answer to their scientific question?
[00:48:57] Sharla: So, part of it is, we’ve done some where we’ve given the raw data. I think we’ve only got one… No, no, no, we have a couple of customers now that ask for the raw data, or at least they ask for the Imaris file. So we can take the… We have our raw data [crosstalk] converted into-
[00:49:17] Aleksandra: An Imaris file is the 3D-rendered image of-
[00:49:20] Sharla: It’s the 3D rendered of it. So what we do is, if we do a tile scan, we’ll check and make sure that the stitch is properly aligned, because sometimes they’re just slightly off in how these softwares align that. And then we’ll provide the merged image, so it’s not as much data. So we get, “Here’s the field of view and then here’s the merge,” so it’s usually almost twice as much data on our side.
[00:49:45] But then on their side, either we’re working with them to understand what they want, and then we’re trying to provide that information to them, or we’re exporting and providing them the file, and they’ve decided, “We want to be able to figure this out and what’s going on.” Imaris has like a free viewer, so we always recommend that, “Hey, we’ll send you the files. This is a great way for you to do it. Make sure that your computer has the minimum requirements in order for you to see it.”
[00:50:12] Aleksandra: At least a gaming laptop, otherwise no?
[00:50:14] Sharla: Yeah. Yeah. Yeah, otherwise you’re probably going to keep crashing your computer.
[00:50:20] Aleksandra: Wow. So we have this, as digital pathologists, we already have this technology challenge of images being big and, I don’t know, bandwidth not being wide enough. And this is like the next level. This is multiple markers, and this is 3D. So I see we are definitely restricted by the current technology to unlock all the insights that could be unlocked with this technology.
[00:50:49] Sharla: Yeah. I mean, it’s a challenge for us, too. Doing this for lab services means that the amount of data that we’re acquiring now is significantly more than what we were doing when we were working on it from an internal perspective. And then it’s finding that balance, too, right? Where, what’s going to be the appropriate step size for you to still get this, and managing the acquisition time to even acquire the images to begin with. And then, what is your ultimate end goal? Most people on their side, they’re like, “Okay, we’re going to figure this out internally,” and that’s their right, because our thing is, “It’s your data. You paid for it.”
[00:51:31] Aleksandra: Yeah.
[00:51:33] Sharla: Sometimes it might be easier for you to be able to see it, but we want to make sure that you actually can see it, and I think that that’s… You can always do… BigDataViewer works well for these larger ones, but it’ll only show you the 2D slicing of it. And they also don’t have colors, so that might be a little bit tough. But you’ve got Imaris, some people that have invested in the license, because it’s just beneficial that way. And then you’ve always got Fiji image viewer, too, but again, trade-offs. Whatever works for you. We’re definitely not here to hold you back, but we can always export it to like a TIFF-OME so if that they have some other preferred imaging platform, then you can just incorporate it into their… And then you get to figure it out.
[00:52:18] Aleksandra: Yeah, I guess, I mean, that’s the next thing.
[00:52:22] Sharla: Well, we all want ownership of it, right? So we can say… Because you never know what’s going to pop up.
[00:52:28] Aleksandra: Mm-hmm [affirmative].
[00:52:29] Sharla: And sometimes it’s faster if you do it yourself, right? Because you don’t have to spend time explaining what it is that you’re after. But depending on what you have available to you, or if just the size of these files is just beyond your management, then it makes more sense to say, “These are the things that I’m looking after,” and then we go after that and we try to make that happen for you. We’ve definitely had some where you go back in and do another… Focus on another region of interest here and provide that.
[00:52:58] Aleksandra: But, good, I think especially in the new methods, this back-and-forth with your customer and this trying to understand the capabilities on one end of your company and of the CLARITY technology, and then what people are actually looking for before even starting to work on it, is crucial. Because on one end, it does let you get into a rabbit hole, and you do a lot more targeted analysis, which I think, with this amount of information, is crucial, because otherwise you’re just going to drown in data.
[00:53:35] Sharla: Oh, that’s true. It’s absolutely true, and what we’ve also learned in doing this is that there’s part of us… Part of it is us just educating people on either what tissue clearing is or maybe how CLARITY is different from other methods. But then what you’re getting when you see it, because we show videos, we try to walk you through it, but sometimes you actually need to see it as it directly applies to your samples, before it’s like, “Oh, okay. All right, now that makes sense.”
[00:54:09] Aleksandra: Yeah.
[00:54:09] Sharla: We have a couple of local customers in the Bay Area that have come by, either to see the data in real time on our computer, or maybe we’re doing like a virtual one where they’re virtually looking at the desktop, or they physically want to come in and see what their cleared sample looks like. And that’s always an interesting experience too, because when it’s new stuff, we’re like, “Oh, yeah, this is cool. Oh, look at that. You can see all of these things.” And in other cases, you get to see other people experience it for the first time, and you’re just like, “Okay, it’s not just me. This is really cool.” It’s really exciting for us.
[00:54:45] Aleksandra: So you’re in the Bay Area, you’re in California?
[00:54:49] Sharla: We’re in the San Francisco Bay Area. So we’re in Sunnyvale, which is south part of the Bay, but yeah. So when the wildfires aren’t getting to us-
[00:54:58] Aleksandra: Wow.
[00:54:58] Sharla: … we’re in great shape.
[00:55:00] Aleksandra: Yeah.
[00:55:00] Sharla: But it’s a good area. We’ve actually had customers kind of across the country, and a couple of international ones as well that we’ve worked with, but we’ve got a wide gamut, which really made it apparent that we were on the right track with, people want to be able to do CLARITY and they want to be able to apply it to a multitude of tissues.
[00:55:25] It’s not even just about cancer anymore. People have some wonderful ideas and some great research that they’re pursuing, and I think we’re just happy to be a part of it. And when we hit those new ones for us, we have no shame in saying like, “We haven’t done this before, but we’ll give it a try.” We did a mouse mandible once, where that was on the table, so that was exciting. Beautiful images too.
[00:55:48] Aleksandra: Yeah. I need to link some, so that our listeners can have a look at what this is. So, what’s ClearLight’s mission? You say you do a lot of stuff for many different researchers, many different tissues, but what’s the mission, and why is your mission important?
[00:56:07] Sharla: So, our goal is to help researchers see more biology, and these are for looking with in the tumor microenvironment, no matter what your tissue microenvironment is. And we want to develop these technologies to better improve the diagnostic, the prognostic and predictive treatments of diseases so that it helps all live a healthier life.
[00:56:35] We think this is important because… So 3D IHC offers these advantages to researchers where you actually can understand what’s going on in real time by locking these tissues into place. And then by doing that, we can improve the chances that they’re going to be able to develop that next level of therapy that can then result in disease prevention.
[00:56:59] Aleksandra: So, from the… I don’t know, is there a competition from the technology standpoint?
[00:57:06] Sharla: I think direct apples-to-apples competition, we don’t have any. There are companies that offer tissue clearing services as a lab services. So in that regard, they might be competition, but none of them are offering CLARITY as their lab services. And then there are companies that have software where they’re doing spatial biology, and in this case it’s really kind of an XY, “Here’s a line and we’re measuring this distance.”
[00:57:36] Aleksandra: Okay.
[00:57:36] Sharla: But what we’re working on is not just that line, we’re doing the XYZ position. So, “Here’s your exact location of this one compared to the XYZ position of that.” So it’s distance and space, for both of those. So combining those together, we don’t have direct competition, but there are people… So if you’ve got the circle of tissue clearing and the circle of spatial analysis, ClearLight’s the only one in the overlap. I can’t think of the name of that… I’m drawing on blank on the intersection of the circle right now. But ClearLight’s in the middle, and then we’ve got some competition in those [crosstalk] .
[00:58:13] Aleksandra: The edges of… Mm-hmm [affirmative]. So, you said you do incorporate artificial intelligence for image analysis. This is definitely big data. Are you doing some artificial intelligence to create insights of your data, or what role is AI playing at ClearLight?
[00:58:33] Sharla: So, for us, it’s all living in the analysis software where we’re doing this 3D quantitation that’s going on, and from feeding it so that all of the information and individual annotations continue to help it improve its response, to then being able to apply this to different tissues. So, often the approach right now is, you focus on one specific disease type, and then you have another specific disease type. So, for us, one of the key aspects that we are working on trying to do this, in addition to that Z-stack and adding in the Z dimension for this, is making this generally and broadly applicable where the software and the algorithms continue to grow based off of all of the additional annotations that are fed into it.
[00:59:29] So that lives pretty much with the software part of it. Everything else on the front end is more biological, molecular biological type aspect for us right there, and that big gap of doing 2D versus 3D analysis is where we continue to work hard, so that when we finally put it out there, that it’s relevant, and that it’s useful, and that it’s a solid starting point. The same way we took the approach for lab services, we want to make sure that before we say, “Hey, we can do this,” that we actually can, and we have a robust methodology for how it’s approached.
[01:00:09] Aleksandra: I think it’s fascinating. It’s really like… I don’t know, the method blows my mind, and you talked a lot about it, but it just is a lot to digest, I think.
[01:00:22] Sharla: Yeah. I’m still digesting some of this, years later.
[01:00:27] Aleksandra: But thanks so much for presenting this. Definitely, this is the next thing to go to in immune oncology and targeted therapies. So there is no doubt about this. Spatial is already happening on 2D, and then the next step is spatial on 3D. Before we go, tell the listeners, where can we find more information about you?
[01:00:50] Sharla: Absolutely. You can find more information about ClearLight Biotechnologies at www.clearlightbio.com. We have a wealth of resources, from links to our YouTube channel, with videos. We have-
[01:01:05] Aleksandra: Oh, yeah.
[01:01:06] Sharla: … infographics that give you the overview of the CLARITY process. There are some infographics about if you wanted to purchase the reagent kits or the lab services. Just a general fact sheet about the company. We try to put it all out there, and then we also have a chat function that’s available so that if there are any questions that you might have, or if you actually want to speak to somebody, that’s also available to you on the website as well.
[01:01:32] Aleksandra: Mm-hmm [affirmative]. I’m going to definitely link this website in the show notes as well. Thank you so much for taking your time-
[01:01:39] Sharla: Thank you.
[01:01:39] Aleksandra: … and telling us about CLARITY and ClearLight.
[01:01:42] Sharla: I appreciate this.
[01:01:42] Aleksandra: Have a great day.
[01:01:45] Sharla: You do the same thing.
